Search

Leather products are widely used across various fields permeating every aspect of daily life. However during the production of synthetic leather fabrics defects are inevitable directly affecting the quality and price of leather products. Early identification of these defects in the production process can prevent further production losses. Nevertheless the high local similarity of defects on leather fabric surfaces causes significant similarities between different types of defects leading to poor detection results. To address this issue the article proposed an end-to-end defect detection method for leather fabric surfaces achieving finer granularity in distinguishing leather defects. To address the high similarity between defect classes this paper introduced an adaptive convolutional attention ACA module. This module comprises channel attention and spatial attention integrating the channel and spatial attention information through a residual structure to generate more discriminative features. Two different-sized convolutional kernels in the spatial attention work in concert to effectively enhance the differences between defects focusing attention on target features and thus reducing the interference response to background information. To amplify the differences between leather fabric defect classes ACA was incorporated into the backbone network to enhance the semantic feature representation capabilities. This integration not only improves the network's ability to differentiate between defect types but also ensures more accurate detection outcomes. Then the article designed a feature pyramid network based on adaptive convolutional attention AC-FPN to improve multi-scale fusion. By leveraging the feature information enhanced by ACA the network enables the flow of information between different scales allowing for finer differentiation between defects and background. Such enhancement significantly improves the detection capability of defects at different scales achieving finer granularity in leather defect differentiation. The multi-scale fusion process ensures that defects of various sizes and shapes are accurately detected regardless of their scale contributing to a more robust detection system. Finally the traditional detection head was replaced with the side-aware boundary localization SABL detection head enabling precise localization of leather fabric defects. The SABL detection head is specifically designed to enhance the accuracy of defect localization by focusing on the boundaries of defects ensuring that even the smallest and most subtle defects are accurately identified and localized. This replacement is crucial for improving the overall precision of the defect detection system making it more reliable for practical applications in leather fabric production. The article validated the proposed method using a self-constructed leather fabric dataset and compared it with different methods. Experimental results demonstrate that the proposed method achieves better performance in distinguishing between different defect types with similar appearances. Compared to other methods this method exhibits superior detection accuracy across various defect types with AP AP50 and AP75 evaluation metrics reaching 83. 4 89. 7 and 85. 6 respectively. This provides a new perspective for automated surface defect detection of leather fabrics. The improved accuracy metrics indicate that the proposed method is highly effective in identifying and classifying defects with significant improvement over existing methods. The proposed defect detection method for leather fabrics demonstrates better performance compared to other methods offering new feasibility for defect detection. Despite the advantages mentioned above the use of a dataset primarily comprising grayscale images may reduce the ability to extract effective information for some colored leather defects. In future research color cameras can be used to capture images of leather defects and incorporate color information to distinguish some leather fabric surface defects. Additionally exploring advanced image processing techniques and integrating them with the current approach could further enhance the defect detection capabilities so as to make the system more versatile and applicable to a wider range of leather products. [ABSTRACT FROM AUTHOR]

Seed purity assurance is an important aspect of maintaining the quality standards of wheat seeds. It relies significantly on quality parameters, like varietal classification and geographical origin identification. Hyperspectral imaging (HSI) has emerged as an advanced nondestructive technique to determine various quality parameters. In recent years, several studies have utilized HSI for varietal classification, although a limited number of varieties were considered. Additionally, no attention has been paid to determining the geographical origin of wheat seeds. To address these gaps, two separate experiments were performed for varietal classification and geographical origin identification. The seeds from 96 varieties grown across 5 different agricultural regions in India were collected. Hyperspectral images of wheat seeds were acquired in the wavelength ranging 900–1700 nm. The spectral reflectance values were obtained from the region of interest (ROI) corresponding to each seed. Subsequently, the deep learning models (convolutional neural networks [CNNs]) were established and compared with two conventional algorithms, including support vector machines (SVMs) and K‐nearest neighbors (KNNs). The experimental results indicated that the proposed CNN models outperformed the SVM and KNN models, achieving an overall accuracy of 94.88% and 99.02% for varietal classification and geographical origin identification, respectively. These results demonstrate that HSI combined with deep learning has the potential to accurately classify a large number of wheat varieties. Moreover, HSI can be used to precisely identify the geographical origins of wheat seeds. This study provides an accurate and nondestructive method that can assist in breeding, quality evaluation, and the development of high‐quality wheat seeds. This study explores the use of hyperspectral imaging (HSI) combined with deep learning for wheat seed quality assessment. We conducted experiments on 96 wheat varieties from five Indian geographical regions, utilizing convolutional neural networks (CNNs) for varietal classification and geographical origin identification. Results showed CNNs outperforming support vector machines (SVMs) and K‐nearest neighbors (KNNs), with accuracies of 94.88% and 99.02%, respectively. Our findings highlight the potential of HSI as an accurate, non‐destructive method for wheat seed quality evaluation. [ABSTRACT FROM AUTHOR]

This paper introduces a novel normalization operation, Channel Modulus Normalization (CMN), to enhance the image classification performance of traditional convolutional neural networks (CNNs) by leveraging latent feature characteristics across all levels. CMN is seamlessly integrated into existing CNN architectures without using additional trainable parameters. Following convolution operations at every level, CMN is applied to each feature map, utilizing a normalization adjustment factor to mitigate the influence of feature modulus length. This operation pays more attention to the change characteristics of the image details for classification. Experiments on six datasets such as CIFAR10, CIFAR100, FaceScrub, Tiny ImageNet, ImageNet (100), and ImageNet (1000), show that the channel modulus normalization operation can effectively improve the classification accuracy of the datasets above. Taking ResNet50 as an example, it is increased by 0.57%, 3.17%, 1.11%, 2.28%, 0.61% and 0.14% respectively. At the same time, the normalization operation will not increase the number of parameters of the model, and the convergence performance during training is better, which will only increase a small amount of calculation. The implementation code is available at . [ABSTRACT FROM AUTHOR]

Background: The urgent need for affordable and rapid detection methodologies for foodborne pathogens, particularly Escherichia coli (E. coli), highlights the importance of developing efficient and widely accessible diagnostic systems. Dark field microscopy, although effective, requires specific isolation of the target bacteria which can be hindered by the high cost of producing specialized antibodies. Alternatively, M13 bacteriophage, which naturally targets E. coli, offers a cost-efficient option with well-established techniques for its display and modification. Nevertheless, its filamentous structure with a large length-diameter ratio contributes to nonspecific binding and low separation efficiency, posing significant challenges. Consequently, refining M13 phage methodologies and their integration with advanced microscopy techniques stands as a critical pathway to improve detection specificity and efficiency in food safety diagnostics. Methods: We employed a dual-plasmid strategy to generate a truncated M13 phage (tM13). This engineered tM13 incorporates two key genetic modifications: a partial mutation at the N-terminus of pIII and biotinylation at the hydrophobic end of pVIII. These alterations enable efficient attachment of tM13 to diverse E. coli strains, facilitating rapid magnetic separation. For detection, we additionally implemented a convolutional neural network (CNN)-based algorithm for precise identification and quantification of bacterial cells using dark field microscopy. Results: The results obtained from spike-in and clinical sample analyses demonstrated the accuracy, high sensitivity (with a detection limit of 10 CFU/μL), and time-saving nature (30 min) of our tM13-based immunomagnetic enrichment approach combined with AI-enabled analytics, thereby supporting its potential to facilitate the identification of diverse E. coli strains in complex samples. Conclusion: The study established a rapid and accurate detection strategy for E. coli utilizing truncated M13 phages as capture probes, along with a dark field microscopy detection platform that integrates an image processing model and convolutional neural network. [ABSTRACT FROM AUTHOR]

Hyperspectral images (HSIs) are rich in spatial and spectral information, making them crucial for accurate classification. While existing convolutional neural network (CNN) methods have shown promise in HSI classification, they primarily focus on local features and struggle with establishing long-range dependencies. To address these limitations, we propose a Local-Global Feature Fusion Network (LGFFN). LGFFN effectively learns joint local-global information from HSIs. Firstly, we employ a group-general convolution parallel module (GGCPM) to extract both local and global spectral features. Next, we utilize parallel branches: a CNN branch for capturing fine-grained spatial information and an improved transformer encoder (ITE) branch for capturing long-distance spatial-spectral features and exploiting semantic information. Finally, we elegantly integrate the diverse features obtained from these branches to obtain classification results. We validate our method on four datasets and compare it with six state-of-the-art hyperspectral classification methods. Our experimental results demonstrate that LGFFN achieves high classification accuracy, outperforming existing methods. [ABSTRACT FROM AUTHOR]

Recently, the convolutional neural network (CNN) has gained widespread adoption in the hyperspectral image (HSI) classification owing to its remarkable feature extraction capability. However, the fixed acceptance domain of CNN restricts it to Euclidean image data only, making it difficult to capture complex information in hyperspectral data. To overcome this problem, much attention has been paid to the graph attention network (GAT), which can effectively model graph structure and capture complex dependencies between nodes. However, GAT usually acts on superpixel nodes, which may lead to the loss of pixel-level information. To better integrate the advantages of both, we propose a CNN-assisted multi-hop graph attention network (CMGAT) for HSI classification. Specifically, a parallel dual-branch architecture is first constructed to simultaneously capture spectral-spatial features from hyperspectral data at the superpixel and pixel levels using GAT and CNN, respectively. On this basis, the multi-hop and multi-scale mechanisms are further employed to construct a multi-hop GAT module and a multi-scale CNN module to capture diverse feature information. Secondly, an attention module is cascaded before the multi-scale CNN module to improve classification performance. Eventually, the output information from the two branches is weighted and fused to produce the classification result. We performed experiments on four benchmark HSI datasets, including Indian Pines (IP), University of Pavia (UP), Salinas Valley (SV) and WHU-Hi-LongKou (LK). The results demonstrate that the proposed method outperforms several deep learning methods, achieving overall accuracies of 95.67%, 99.04%, 99.55% and 99.51%, respectively, even with fewer training samples. [ABSTRACT FROM AUTHOR]

Hyperspectral image classification (HSI) is a vital aspect of remote sensing technology. However, the high-dimensional nature of HSI and the continuous spectral bands significantly impact its classification performance. In this paper, we propose a classification method called MBSFC for HSI classification. MBSFC utilizes multi-branch and spectral feature conversion (SFC) techniques to eliminate redundant information and extract valuable features. Firstly, to preserve spectral-spatial information, the SFC block is introduced to eliminate redundant information through up-sampling. Subsequently, the feature maps from the SFC block and the spectral bands reduced by a factor of four are fed into three branches: the spectral branch, spatial-X branch, and spatial-Y branch. Each branch employs a 3D-CNN-based dense block and attention mechanism to extract useful features. Finally, the obtained features from the three branches are fused for HSI classification. To cater to different application scenarios, we divide MBSFC into three models with different network structures: MBSFC-s, MBSFC-m, and MBSFC-l. The MBSFC-l model, with a three-branch structure, achieves the best performance. The three models differ in the number of branches in the feature extraction. Experimental results on four publicly available hyperspectral datasets show that MBSFC achieves competitive results with small samples compared to other state-of-the-art methods. The code is available at https://github.com/TeresaTing/MBSFC. [ABSTRACT FROM AUTHOR]

Nowadays, deploying an intelligent vehicle detection system (IVDS) in diverse traffic is a work priority. It provides real-time traffic information with vehicle counts and types of vehicles. IVDS deployment in diverse traffic is challenging because different vehicle classes occlude each other on the road. In recent years, convolutional neural network (CNN) based deep learning (DL) methods have attained incredible progress in implementing IVDS. However, most CNN-based DL methods do not include diverse traffic conditions in Asian countries. Also, due to existing feature extraction backbones, they cannot accurately detect multi-scale vehicles. This work proposes an advanced visual computing deep learning (AVCDL) method with a vast labeled vehicle dataset to detect vehicles in diverse traffic. It includes an ensemble backbone and an improved multi-stage vehicle detection head (MSVDH). An ensemble CNN backbone extracts the vehicle features and combines them on a single channel via a feature concatenation. The final detection is carried out by an improved MSVDH that classifies the target vehicles. The proposed method is examined, tested, and evaluated using traffic statistics. It is contrasted with current cutting-edge vehicle detection techniques. It achieves 86.32% mean average precision (mAP) on self-collected diverse traffic labeled dataset (DTLD) and 86.17% mAP on KITTI. Moreover, the real-time performance is validated with NVIDIA Jetson Tx2 and Nano boards. It achieves 15 frames per second (FPS) on Jetson Tx2 and 7 FPS on Jetson Nano. [ABSTRACT FROM AUTHOR]

To improve slope reliability calculations and address high-nonlinearity in random fields, an AI algorithm, namely Convolutional Neural Network (CNN) with asymmetric convolution is introduced. The method accounts for the interdependence and auto-correlation of soil material and uses Python-based secondary development in ABAQUS Version 6.14 to improve computational efficiency and user-friendliness in finite element simulations. A Cholesky decomposition-based centroid point method is used for random fields to simplify computation. Additionally, an asymmetric convolution-based CNN surrogate model replaces finite element simulations to address challenges such as parameter correlations and random field discretization for improved analysis efficiency. The methodology uses random field samples and safety factors as inputs and outputs for training, which improves predictability and addressing high-dimensional issues. Its effectiveness is demonstrated through case studies involving single-layer undrained saturated clay slopes and double-layer cohesive soil slopes. The results demonstrate the effectiveness of the CNN approach that utilizes asymmetric convolution, with outcomes closely resembling those obtained through finite element simulation. This method demonstrates a 95.8% improvement in time efficiency compared to software-based calculations and a 93.5% enhancement over batch calculations using ABAQUS. These results confirm the effectiveness of the introduced reliability analysis method and the ability to provide accurate results while significantly boosting computational efficiency. [ABSTRACT FROM AUTHOR]

To investigate the impact of diverse multivariate mixing excitation conditions on the hysteresis loop of ferromagnetic materials, this study initially constructs a magnetic performance testing system for electrical steel. This system is capable of generating mixed excitation utilizing a standard Epstein square‐circle setup. Subsequently, the study measures the magnetic properties of the oriented silicon steel sheets at various mixing AC frequencies of the hysteresis loop data. Secondly, a hybrid network model integrating a convolutional neural network (CNN) and a bi‐directional long short‐term memory network (BiGRU), augmented with an attention mechanism (AM), is proposed and utilized for predicting the hysteresis properties of oriented silicon steel wafers subjected to compound mixed‐frequency excitation. The model utilizes CNN to extract high‐dimensional data features reflecting the hysteresis characteristics of the loop, BiGRU to capture the temporal evolution patterns of the key feature vectors, an AM to weigh the feature parameters and emphasize the key features, and a Bayesian optimization (BO) algorithm based on neural network hyperparameters for automatic selection, enhancing prediction accuracy. In comparison with experimental observations, the method accurately predicts material hysteresis properties under non‐sinusoidal complex excitation conditions, outperforming existing deep‐learning network models. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Variability in the inherently dynamic nature of additive manufacturing introduces imperfections that hinder the commercialization of new materials. Binder jetting produces ceramic and metallic parts, but low green densities and spreading anomalies reduce the predictability and processability of resulting geometries. In situ feedback presents a method for robust evaluation of spreading anomalies, reducing the number of required builds to refine processing parameters in a multivariate space. In this study, we report layer-wise powder bed semantic segmentation for the first time with a visually light ceramic powder, alumina, or Al2O3, leveraging an image analysis software to rapidly segment optical images acquired during the additive manufacturing process. Using preexisting image analysis tools allowed for rapid analysis of 316 stainless steel and alumina powders with small data sets by providing an accessible framework for implementing neural networks. Models trained on five build layers for each material to classify base powder, parts, streaking, short spreading, and bumps from recoater friction with testing categorical accuracies greater than 90%. Lower model performance accompanied the more subtle spreading features present in the white alumina compared to the darker steel. Applications of models to new builds demonstrated repeatability with the resulting models, and trends in classified pixels reflected corrections made to processing parameters. Through the development of robust analysis techniques and feedback for new materials, parameters can be corrected as builds progress. [ABSTRACT FROM AUTHOR]

• We designed an intelligent two-layer deep learning (DL) model that predict the promoter region in the first stage and their functional types in second stage respectively. • We captured the DNA encoded patterns using Word2Vec algorithm and analyzed the visual impact of biological features using t-distributed stochastic neighbor embedding method. • We enhanced the overall prediction performance of promoters and their functional types (as weak and strong prompters) on multiple datasets. Promoters, which are short (50–1500 base-pair) in DNA regions, have emerged to play a critical role in the regulation of gene transcription. Numerous dangerous diseases, likewise cancer, cardiovascular, and inflammatory bowel diseases, are caused by genetic variations in promoters. Consequently, the correct identification and characterization of promoters are significant for the discovery of drugs. However, experimental approaches to recognizing promoters and their strengths are challenging in terms of cost, time, and resources. Therefore, computational techniques are highly desirable for the correct characterization of promoters from unannotated genomic data. Here, we designed a powerful bi-layer deep-learning based predictor named "PROCABLES", which discriminates DNA samples as promoters in the first-phase and strong or weak promoters in the second-phase respectively. The proposed method utilizes five distinct features, such as word2vec, k-spaced nucleotide pairs, trinucleotide propensity-based features, trinucleotide composition, and electron–ion interaction pseudopotentials, to extract the hidden patterns from the DNA sequence. Afterwards, a stacked framework is formed by integrating a convolutional neural network (CNN) with bidirectional long-short-term memory (LSTM) using multi-view attributes to train the proposed model. The PROCABLES model achieved an accuracy of 0.971 and 0.920 and the MCC 0.940 and 0.840 for the first and second-layer using the ten-fold cross-validation test, respectively. The predicted results anticipate that the proposed PROCABLES protocol outperformed the advanced computational predictors targeting promoters and their types. In summary, this research will provide useful hints for the recognition of large-scale promoters in particular and other DNA problems in general. [ABSTRACT FROM AUTHOR]

• The use of multi-scale convolutional layers can effectively help to identify hidden dependencies between multiple sequences, capture local patterns in the input sequences more flexibly, and extract location-specific features at different levels. • As global response normalization can achieve global feature aggregation, it can help extract more accurate features in the model and fully express the key information of the 6mA site. • The prediction results are better than other models, and a vector of contribution scores is created that clearly explains the prediction mechanism. DNA N6 methyladenine (6mA) plays an important role in many biological processes, and accurately identifying its sites helps one to understand its biological effects more comprehensively. Previous traditional experimental methods are very labor-intensive and traditional machine learning methods also seem to be somewhat insufficient as the database of 6mA methylation groups becomes progressively larger, so we propose a deep learning-based method called multi-scale convolutional model based on global response normalization (CG6mA) to solve the prediction problem of 6mA site. This method is tested with other methods on three different kinds of benchmark datasets, and the results show that our model can get more excellent prediction results. [ABSTRACT FROM AUTHOR]

Many major diseases of the retina often show symptoms of lesions in the fundus of the eye. The extraction of blood vessels from retinal fundus images is essential to assist doctors. Some of the existing methods do not fully extract the detailed features of retinal images or lose some information, making it difficult to accurately segment capillaries located at the edges of the images. In this paper, we propose a multi-scale retinal vessel segmentation network (SCIE_Net) based on skip connection information enhancement. Firstly, the network processes retinal images at multiple scales to achieve network capture of features at different scales. Secondly, the feature aggregation module is proposed to aggregate the rich information of the shallow network. Further, the skip connection information enhancement module is proposed to take into account the detailed features of the shallow layer and the advanced features of the deeper network to avoid the problem of incomplete information interaction between the layers of the network. Finally, SCIE_Net achieves better vessel segmentation performance and results on the publicly available retinal image standard datasets DRIVE, CHASE_DB1, and STARE. [ABSTRACT FROM AUTHOR]

In clinical practice, the morphology of the left atrial appendage (LAA) plays an important role in the selection of LAA closure devices for LAA closure procedures. The morphology determination is influenced by the segmentation results. The LAA occupies only a small part of the entire 3D medical image, and the segmentation results are more likely to be biased towards the background region, making the segmentation of the LAA challenging. In this paper, we propose a lightweight attention mechanism called fusion attention, which imitates human visual behavior. We process the 3D image of the LAA using a method that involves overview observation followed by detailed observation. In the overview observation stage, the image features are pooled along the three dimensions of length, width, and height. The obtained features from the three dimensions are then separately input into the spatial attention and channel attention modules to learn the regions of interest. In the detailed observation stage, the attention results from the previous stage are fused using element-wise multiplication and combined with the original feature map to enhance feature learning. The fusion attention mechanism was evaluated on a left atrial appendage dataset provided by Liaoning Provincial People's Hospital, resulting in an average Dice coefficient of 0.8855. The results indicate that the fusion attention mechanism achieves better segmentation results on 3D images compared to existing lightweight attention mechanisms. [ABSTRACT FROM AUTHOR]

In-situ leaf chlorophyll content (LCC) estimation based on hyperspectral imaging (HSI) is crucial to track the growth status of crops for field management. However, spatial and spectral features of HSI data, suffering from interference of growth dynamic effect and soil, pose the challenge on accuracy and robustness of LCC estimation in several years and growth stages. Therefore, a joint spectral-spatial feature extraction method was proposed by cascade of three-dimensional convolutional neural network (3DCNN) and long short-term memory (LSTM) to reduce the interference for optimising the LCC estimation. Firstly, crop pixels were separated from soil with vegetation index segmentation method. Secondly, when raw images and segmented pixels were input, sensitive bands were selected by random frog (RF bands), and 3DCNN-LSTM was used to extract the joint spectral-spatial features. Finally, models established by RF bands, 3DCNN and 3DCNN-LSTM were compared, and robustness in individual years and stages was validated. Results showed that RF bands and 3DCNN obtained R P 2 of 0.76 and 0.84 when not segmented. After segmentation, performance of 3DCNN improved (R P 2 = 0.85) compared to RF bands (R P 2 = 0.80). Spectral-spatial features by 3DCNN reduced the interference of soil. 3DCNN-LSTM without and with segmentation obtained good performance with R P 2 of 0.95 and 0.96, and the proposed method could reduce the image segmentation process. The optimal model achieved R P 2 above 0.93 in individual years (R P 2 = 0.96 in 2021, R P 2 = 0.94 in 2021) and R P 2 in the range of 0.87–0.97 at individual stages. This paper provides a method to track growth variability between soil and crop for the LCC estimation optimisation. • 3DCNN-LSTM method was proposed to extract spectral-spatial feature for LCC estimation. • 3DCNN reduced the interference of soil and crop difference. • 3DCNN-LSTM method can reduce the image segmentation procedure. • LSTM captured growth dynamics and improved robustness in multiple years and stages. [ABSTRACT FROM AUTHOR]

Photoacoustic imaging (PAI) is a promising transcranial imaging technique. However, the distortion of photoacoustic signals induced by the skull significantly influences its imaging quality. We aimed to use deep learning for removing artifacts in PAI. In this study, we propose a polarized self-attention dense U-Net, termed PSAD-UNet, to correct the distortion and accurately recover imaged objects beneath bone plates. To evaluate the performance of the proposed method, a series of experiments was performed using a custom-built PAI system. The experimental results showed that the proposed PSAD-UNet method could effectively implement transcranial PAI through a one- or two-layer bone plate. Compared with the conventional delay-and-sum and classical U-Net methods, PSAD-UNet can diminish the influence of bone plates and provide high-quality PAI results in terms of structural similarity and peak signal-to-noise ratio. The 3-D experimental results further confirm the feasibility of PSAD-UNet in 3-D transcranial imaging. PSAD-UNet paves the way for implementing transcranial PAI with high imaging accuracy, which reveals broad application prospects in preclinical and clinical fields. [ABSTRACT FROM AUTHOR]

Flooding disaster causes huge impacts on the socio-economic world. In the inundated area, some geo-referenced images are shared through some media posts, which assist in providing alertness to the critical volunteers and managing the financial loss crisis. In this work, the Adaptive Billiards-Inspired Optimization (A-BIO) and Optimized Ensemble-learning-based detection (OED) with map reduce framework is proposed for flood disaster detection. Initially, the big data is gathered and processed for detection. During the map phase, data preprocessing is performed to enhance the performance of the data, which helps in removing the noise or unwanted attributes. Furthermore, the reduction phase can be done through weighted feature selection, where the features to be selected and the weight is optimized through A-BIO, which assists in getting the most significant features for improving the performance and reducing the complexity of the designed model. Finally, OED is performed by a set of classifiers like Convolutional Neural Networks, Adaboost, XGBoost, Long Short-Term Memory, and Deep Neural Networks, where the parameters of ensemble learning classifiers are optimized by A-BIO algorithm. Finally, through the performance analysis, this detection model can provide high accuracy and better detection performance to avoid the huge impacts of flood disasters. [ABSTRACT FROM AUTHOR]

Several traditional methods were tested on standard datasets to evaluate client emotions transmitted via internet portals. Customers, on the other hand, continue to have difficulty obtaining aspect-oriented viewpoints voiced by other customers, and the accuracy of the current model is insufficient. The suggested Aspect-Based Sentimental Analysis (ABSA) starts with pre-processing, which includes "stop word and punctuation removal, lower case conversion, and stemming." Aspect extraction, which entails dividing the nouns and adjectives, as well as verbs and adverbs, is the following step. The weighted polarity features from the "Vader sentiment intensity analyzer, as well as the word2vector and Term Frequency-Inverse Document Frequency (TF-IDF)" are concatenated. OIDL stands for Optimized Integrated Deep Learning, which combines two types of deep learners. The first is the combination of concatenated features with "Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN)," while the second is the combination of concatenated features with RNN. The Improved Coyote Optimization Algorithm (ICOA) improves both deep learners, and the conclusion of sentiment analysis result is considered both models. Thus, the suggested model surpasses standard methodologies regarding precision and accuracy, according to the results of the experiments. [ABSTRACT FROM AUTHOR]

This paper investigates the nuanced characteristics of the spectral envelope attributes due to vocal-tract resonance structure and fine-level excitation source features within short-term Fourier transform (STFT) magnitude spectra for the assessment of dysarthria. The single-channel convolutional neural network (CNN) employing time-frequency representations such as STFT spectrogram (STFT-SPEC) and Mel-spectrogram (MEL-SPEC) does not ensure capture of the source and system information simultaneously due to the filtering operation using a fixed-size filter. Building upon this observation, this study first explores the significance of convolution filter size in the context of the CNN-based automated dysarthric assessment system. An approach is then introduced to effectively capture resonance structure and fine-level features through a multi-channel CNN. In the proposed approach, the STFT-SPEC is decomposed using a one-level discrete wavelet transform (DWT) to separate the slow-varying spectral structure and fine-level features. The resulting decomposed coefficients in four directions are taken as the inputs to multi-channel CNN to capture the source and system features by employing different sizes of convolution filters. The experimental results conducted on the UA-speech corpus validate the efficacy of the proposed approach utilizing multi-channel CNN. The proposed approach demonstrates the notable enhancement in accuracy and F1 score (60.86% and 48.52%) compared to a single-channel CNN using STFT-SPEC (46.45% and 40.97%), MEL-SPEC (48.86% and 38.20%), and MEL-SPEC appended with delta and delta-delta coefficients (52.40% and 42.84%) for assessment of dysarthria in a speaker-independent and text-independent mode. [ABSTRACT FROM AUTHOR]

Background: Being an image-based optical technique for full-field deformation measurements, the ultimate purpose of digital image correlation (DIC) is to realize accurate, precise and pixel-wise displacement/strain measurements in a full-automatic manner without users' inputs. Objective: In this work, we propose a task-optimized neural network, called RAFT-DIC, to achieve user-independent, accurate and pixel-wise displacement field measurements. Methods: RAFT-DIC is based on the state-of-the-art optical flow architecture: Recurrent All-Pairs Field Transforms (RAFT). We make two targeted improvements that fundamentally enhanced its measurement accuracy and generalization performance. Firstly, we remove all the down-sampling operations in the encode module to improve the perception of spatial information, and reduce the number of pyramid levels of the correlation layer to increase the small displacement accuracy. By building the correlation layer to compute the similarity of pixel pairs, and iteratively updating the displacement field through a recurrent unit, RAFT-DIC introduces the prior information of DIC measurement to guide the displacement estimation with high accuracy. Secondly, we develop a novel dataset generation method to synthesize customized speckle patterns and diverse displacement fields, which facilitate the construction of a robust and adaptable dataset to improve the network generalization. Results: Both simulated and real experimental results demonstrate that the accuracy of the proposed method is approximately an order of magnitude higher than pervious deep learning-based DIC (DL-DIC). Conclusions: The proposed RAFT-DIC shows higher accuracy as well as stronger practicality and cross-dataset generalization performance over existing DL-DIC methods, and is expected to be a new standard architecture for DL-DIC. [ABSTRACT FROM AUTHOR]

Model performance has been significantly enhanced by channel attention. The average pooling procedure creates skewness, lowering the performance of the network architecture. In the channel attention approach, average pooling is used to collect feature information to provide representative values. By leveraging the central limit theorem, we hypothesize that the strip-shaped average pooling operation will generate a one-dimensional tensor by considering the spatial position information of the feature map. The resulting tensor, obtained through average pooling, serves as the representative value for the features, mitigating skewness during the process. By incorporating the concept of the central limit theorem into the channel attention operation process, this study introduces a novel attention mechanism known as the "Central Attention Mechanism (CAM)." Instead of directly using average pooling to generate channel representative values, the central attention approach employs star-stripe average pooling to normalize multiple feature representative values into a single representative value. In this way, strip-shaped average pooling can be utilized to collect data and generate a one-dimensional tensor, while star-stripe average pooling can provide feature representative values based on different spatial directions. To generate channel attention for the complementary input features, the activation of the feature representation value is performed for each channel. Our attention approach is flexible and can be seamlessly incorporated into various traditional network structures. Through rigorous testing, we demonstrate the effectiveness of our attention strategy, which can be applied to a wide range of computer vision applications and outperforms previous attention techniques. [ABSTRACT FROM AUTHOR]

Diabetes retinopathy (DR) is one of the complications of diabetes. Early diagnosis of retinopathy is helpful to avoid vision loss or blindness. The difficulty of this task lies in the significant differences in the size and shape of lesions between different DR samples, with a higher proportion of small lesions. We propose a new multi-disease segmentation method based on UNet++ to improve the segmentation accuracy of DR lesions. We chose Resnet50 as the backbone network and introduced a new hybrid residual module to replace the original residual module. At the same time, to compensate for the loss of information in DR small lesions during the feature extraction process, we introduce the Across Feature Map Attention (AFMA) is an auxiliary branch that enhances the segmentation accuracy of small-scale lesions. Finally, in response to the difficulty in extracting DR lesions in shallow models, the model abandoned the deep supervision structure in UNet++. In addition, we use a weighted mixed loss function to train the model. We conducted experiments on IDRID and DDR public datasets, simultaneously segmenting four typical DR lesions. The results on intersection over union (IOU) and dice similarity coefficient (Dice) showed that our method achieved competitive performance compared to other research methods. [ABSTRACT FROM AUTHOR]

Copyright of Lubrication Engineering (0254-0150) is the property of Editorial Office of LUBRICATION ENGINEERING and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Lubrication Engineering (0254-0150) is the property of Editorial Office of LUBRICATION ENGINEERING and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of South China University of Technology (Natural Science Edition) is the property of South China University of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In recent years, attention mechanisms have greatly contributed to the representation ability of deep convolutional neural networks, especially the attention module in convolutional neural networks. This paper proposes an operational unit called Map-and-Acquisition (MA) to compute the attention of different feature vectors in convolutional neural networks based on simple mathematical principles. Inspired by the convolution operation, MA extends the convolution operator to integrate spatial or channel features in the local receptive field to the global spatial or channel, and only operates on one dimension of the image feature, avoiding a series of problems caused by the coupling of spatial and channel dimensions. The feature response of the channel or spatial domain is adaptively adjusted through the weighting and summing operator to better extract the high-level features of the image. Our method is verified through image classification, object detection, and instance segmentation on the ImageNet-1K and COCO 2017 datasets, and is also evaluated on the small-sample image classification task on the dataset mini-Imagenet. Experimental results show that our method can significantly improve the performance of vision tasks. [ABSTRACT FROM AUTHOR]

With the development of Convolutional Neural Networks (CNNs), there is a growing requirement for their deployment on edge devices. At the same time, Compute-In-Memory (CIM) technology has gained significant attention in edge CNN applications due to its ability to minimize data movement between memory and computing units. However, the deployment of complex deep neural network models on edge devices with restricted hardware resources continues to be challenged by a lack of adequate storage for intermediate layer data. In this article, we propose an optimized JPEG Lossless Compression (JPEG-LS) algorithm that implements serial context parameter updating alongside parallel encoding. This method is designed for the global prediction and efficient compression of intermediate data layers in neural networks employing CIM techniques. The results indicate average compression ratios of 6.44× for VGG16, 3.62× for ResNet34, 1.67× for MobileNetV2, and 2.31× for InceptionV3. Moreover, the implementation achieves a data throughput of 32 bits per cycle at 600 MHz on the TSMC 28 nm, with a hardware cost of 122 K Gate Count. [ABSTRACT FROM AUTHOR]

This paper addresses the practical challenge of detecting tomato plant diseases using a hybrid lightweight model that combines a Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). Traditional image classification models demand substantial computational resources, limiting their practicality. This study aimed to develop a model that can be easily implemented on low-cost IoT devices while maintaining high accuracy with real-world images. The methodology leverages a CNN for extracting high-level image features and an RNN for capturing temporal relationships, thereby enhancing model performance. The proposed model incorporates a Closed-form Continuous-time Neural Network, a lightweight variant of liquid time-constant networks, and integrates Neural Circuit Policy to capture long-term dependencies in image patterns, reducing overfitting. Augmentation techniques such as random rotation and brightness adjustments were applied to the training data to improve generalization. The results demonstrate that the hybrid models outperform their single pre-trained CNN counterparts in both accuracy and computational cost, achieving a 97.15% accuracy on the test set with the proposed model, compared to around 94% for state-of-the-art pre-trained models. This study provides evidence of the effectiveness of hybrid CNN-RNN models in improving accuracy without increasing computational cost and highlights the potential of liquid neural networks in such applications. [ABSTRACT FROM AUTHOR]

This systematic literature review employs the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology to investigate recent applications of explainable AI (XAI) over the past three years. From an initial pool of 664 articles identified through the Web of Science database, 512 peer-reviewed journal articles met the inclusion criteria—namely, being recent, high-quality XAI application articles published in English—and were analyzed in detail. Both qualitative and quantitative statistical techniques were used to analyze the identified articles: qualitatively by summarizing the characteristics of the included studies based on predefined codes, and quantitatively through statistical analysis of the data. These articles were categorized according to their application domains, techniques, and evaluation methods. Health-related applications were particularly prevalent, with a strong focus on cancer diagnosis, COVID-19 management, and medical imaging. Other significant areas of application included environmental and agricultural management, industrial optimization, cybersecurity, finance, transportation, and entertainment. Additionally, emerging applications in law, education, and social care highlight XAI's expanding impact. The review reveals a predominant use of local explanation methods, particularly SHAP and LIME, with SHAP being favored for its stability and mathematical guarantees. However, a critical gap in the evaluation of XAI results is identified, as most studies rely on anecdotal evidence or expert opinion rather than robust quantitative metrics. This underscores the urgent need for standardized evaluation frameworks to ensure the reliability and effectiveness of XAI applications. Future research should focus on developing comprehensive evaluation standards and improving the interpretability and stability of explanations. These advancements are essential for addressing the diverse demands of various application domains while ensuring trust and transparency in AI systems. [ABSTRACT FROM AUTHOR]

Due to the high detection efficiency of the airborne time-domain electromagnetic method, it can quickly collect electromagnetic response data for large area-wide regions, but it also brings great challenges to the inversion interpretation of the data because there are numerous survey data that need to be inverted. Conventional optimal inversion and fast imaging methods still take a long time to obtain conductivity and depth information, which affect the efficiency of real-time data interpretation. In this paper, we present a deep learning inversion method that can be used to solve the fast inversion problem of airborne time-domain electromagnetic data; the method uses a one-dimensional convolutional neural network. The network structure consists of two parts containing different numbers of convolutional and pooling layers. The training sample dataset was generated via two ways of constructing geoelectric models through forward modelling. To check the effectiveness of our deep learning inversion strategy, we tested it on synthetic data and two types of survey data. The experimental results show that this inversion method is effective and that it can be applied to airborne time-domain electromagnetic data collected using different observation systems. The proposed inversion method can obtain better inversion results for both simple and complex stratigraphic structures and requires significantly less computation time compared to conventional optimal inversion methods. [ABSTRACT FROM AUTHOR]

Featured Application: The model proposed in this paper can help stock investors to get high returns on newly listed stocks. Most previous stock investing methods were unable to predict newly listed stocks because they did not have historical data on newly listed stocks. In this paper, we use the Q-learning algorithm based on a convolutional neural network and add sentiment analysis to establish a prediction method for Chinese stock investment tasks. There are 118 companies that are ranked in the Chinese top 150 list for two consecutive years in both 2022 and 2023. We collected all comments under the stock bar of these 118 stocks for each day from 1 January 2022 to 1 July 2024, totaling nearly 10 million comments. There are 90 stocks left after the preprocessing of 118 stocks. We use these 90 stocks as the dataset. The stock's closing price, volume, and comment text data are fed together to the agent, and the trained agent outputs investment behaviors that maximize future returns. We apply the trained model to two test sets that are completely different from the training set and compare it to several other methods. Our proposed method called SADQN-S obtains results of 1.1229 and 1.1054 on the two test sets. SADQN-S obtained higher final total assets than the other methods on both test sets. This shows that our model can help stock investors earn high returns on newly listed stocks. [ABSTRACT FROM AUTHOR]

Near-infrared (NIR) spectroscopy has become a popular technique for assessing food quality due to its advantages over complex chemical analysis methods. However, the application of NIR spectroscopy for evaluating fish quality based on urea content has not been extensively explored. This study investigates the use of NIR spectroscopy in combination with machine learning (ML) techniques to classify fish samples into two safety classes—Safe and Unsafe—based on their urea content. A comprehensive NIR dataset comprising 11,960 spectra collected from eight distinct positions within the fish body was obtained from 299 fish samples of mackerel, tuna, and pompano species. ML experiments were conducted to classify fish samples based on whether their urea content exceeded the permissible limit of 1000 ppm. To address class imbalance and optimize ML models, various data pre-processing and feature extraction techniques, as well as ML algorithms, were explored. The results demonstrated that utilizing NIR data specifically obtained from the outer skin of the stomach yielded superior models for fish safety classification. A feature extraction method employing pre-processed NIR spectra and their first derivatives, combined with an optimized convolutional neural network architecture, outperformed traditional ML classifiers, achieving an accuracy of up to 83.9%. [ABSTRACT FROM AUTHOR]

Climate change affects the water cycle, water resource management, and sustainable socio-economic development. In order to accurately predict climate change in Weifang City, China, this study utilizes multiple data-driven deep learning models. The climate data for 73 years include monthly average air temperature (MAAT), monthly average minimum air temperature (MAMINAT), monthly average maximum air temperature (MAMAXAT), and monthly total precipitation (MP). The different deep learning models include artificial neural network (ANN), recurrent NN (RNN), gate recurrent unit (GRU), long short-term memory neural network (LSTM), deep convolutional NN (CNN), hybrid CNN-GRU, hybrid CNN-LSTM, and hybrid CNN-LSTM-GRU. The CNN-LSTM-GRU for MAAT prediction is the best-performing model compared to other deep learning models with the highest correlation coefficient (R = 0.9879) and lowest root mean square error (RMSE = 1.5347) and mean absolute error (MAE = 1.1830). These results indicate that The hybrid CNN-LSTM-GRU method is a suitable climate prediction model. This deep learning method can also be used for surface water modeling. Climate prediction will help with flood control and water resource management. [ABSTRACT FROM AUTHOR]

The fields of image processing and computer vision have witnessed significant growth due to the proliferation of digital images across diverse domains. Image Segmentation is the fundamental task in digital image processing, finding applications in pivotal areas such as medical imaging, covert communication, autonomous driving, satellite imaging, among others. One particularly intriguing application of image segmentation lies in Reversible Data Hiding (RDH), where the delineation of the main Region of Interest (ROI) and Non-Region of Interest (NROI) using segmentation plays a crucial role for effective data encryption in the images. Over the last two decades, various studies focussed on developing an efficient data hiding approach, which can embed secret data within ROI and NROI part of image while ensuring its quality. A comprehensive survey has been conducted that meticulously examines different segmentation techniques, along with its usage in reversible data hiding. The main objective of this survey is to compare the performance metrics of reversible data hiding after applying different image segmentation techniques. The image segmentation techniques have been categorized systematically into three main classes: i) Traditional segmentation techniques, encompassing a spectrum of approaches like thresholding, region-based and edge detection based techniques, ii) Machine Learning (ML) based approach consisting of Clustering, Support Vector Machine (SVM) and iii) Deep Learning (DL) based technique, propelled by Convolutional Neural Networks (CNNs) that have emerged as a transformative paradigm, revolutionizing segmentation tasks with their ability to learn complex images. The survey finds out that PSNR value of data embedded images is high after applying deep learning based segmentation technique. [ABSTRACT FROM AUTHOR]

Human Interaction Recognition (HIR) was one of the challenging issues in computer vision research due to the involvement of multiple individuals and their mutual interactions within video frames generated from their movements. HIR requires more sophisticated analysis than Human Action Recognition (HAR) since HAR focuses solely on individual activities like walking or running, while HIR involves the interactions between people. This research aims to develop a robust system for recognizing five common human interactions, such as hugging, kicking, pushing, pointing, and no interaction, from video sequences using multiple cameras. In this study, a hybrid Deep Learning (DL) and Machine Learning (ML) model was employed to improve classification accuracy and generalizability. The dataset was collected in an indoor environment with four-channel cameras capturing the five types of interactions among 13 participants. The data was processed using a DL model with a fine-tuned ResNet (Residual Networks) architecture based on 2D Convolutional Neural Network (CNN) layers for feature extraction. Subsequently, machine learning models were trained and utilized for interaction classification using six commonly used ML algorithms, including SVM, KNN, RF, DT, NB, and XGBoost. The results demonstrate a high accuracy of 95.45% in classifying human interactions. The hybrid approach enabled effective learning, resulting in highly accurate performance across different interaction types. Future work will explore more complex scenarios involving multiple individuals based on the application of this architecture. [ABSTRACT FROM AUTHOR]

The surge in connected devices and massive data aggregation has expanded the scale of the Internet of Things (IoT) networks. The proliferation of unknown attacks and related risks, such as zero-day attacks and Distributed Denial of Service (DDoS) attacks triggered by botnets, have resulted in information leakage and property damage. Therefore, developing an efficient and realistic intrusion detection system (IDS) is critical for ensuring IoT network security. In recent years, traditional machine learning techniques have struggled to learn the complex associations between multidimensional features in network traffic, and the excellent performance of deep learning techniques, as an advanced version of machine learning, has led to their widespread application in intrusion detection. In this paper, we propose an Adaptive Particle Swarm Optimization Convolutional Neural Network Squeeze-and-Excitation (APSO-CNN-SE) model for implementing IoT network intrusion detection. A 2D CNN backbone is initially constructed to extract spatial features from network traffic. Subsequently, a squeeze-and-excitation channel attention mechanism is introduced and embedded into the CNN to focus on critical feature channels. Lastly, the weights and biases in the CNN-SE are extracted to initialize the population individuals of the APSO. As the number of iterations increases, the population’s position vector is continuously updated, and the cross-entropy loss function value is minimized to produce the ideal network architecture. We evaluated the models experimentally using binary and multiclassification on the UNSW-NB15 and NSL-KDD datasets, comparing and analyzing the evaluation metrics derived from each model. Compared to the base CNN model, the results demonstrate that APSO-CNNSE enhances the binary classification detection accuracy by 1.84% and 3.53% and the multiclassification detection accuracy by 1.56% and 2.73% on the two datasets, respectively. Additionally, the model outperforms the existing models like DT, KNN, LR, SVM, LSTM, etc., in terms of accuracy and fitting performance. This means that the model can identify potential attacks or anomalies more precisely, improving the overall security and stability of the IoT environment. [ABSTRACT FROM AUTHOR]

Tomato, as an essential food crop, is consumed worldwide, and at the same time, it is susceptible to several diseases that lead to a reduction in tomato yield. Proper diagnosis of tomato diseases is required to increase the output of tomato crops. For this purpose, this paper proposes a tomato plant disease detection algorithm based on Pyramid Scene Parsing Network (PSPNet) and deep learning. First, the training data set is augmented with data to alleviate the data imbalance problem in each category, and then the augmented images are fed into the proposed Mob-PSP network for training. The proposed network utilizes the lightweight MobileNet-V2 model as the feature extraction technique while integrating the PSPNet module to enhance the network’s detection performance. The aim is to effectively extract local and global features from plant disease images, which are being introduced in plant disease detection. This study evaluated the model on the tomato subset of the public data set PlantVillage. The experimental results demonstrate that this algorithm achieves a balance between inference speed and detection accuracy, outperforming other state-of-the-art algorithms. Additionally, compared to the baseline model Inception-V3, the inference speed is improved by 10.73 frames per second, while maintaining an average accuracy of 99.69 % with only 6.5M parameters. [ABSTRACT FROM AUTHOR]

The extraction of building footprints from aerial photos and satellite imagery plays a crucial role in change detection, urban development, and detecting encroachments on agricultural land. Deep neural networks offer the capability of extracting features and provide accurate methods for detecting and extracting building footprints from satellite imagery. Image segmentation, the process of dividing an image into coherent parts, can be accomplished using two types: semantic segmentation and instance segmentation. Convolutional neural networks (CNN) are commonly used for both instance and semantic segmentation tasks. In this paper, we propose a hybrid approach to extracting building footprints from low-resolution satellite imagery using instance segmentation techniques. Our analysis demonstrates that the mask region-based CNN (R-CNN) architecture with a ResNet-34 backbone and PointRend head to improve the bounding-boxes and mask prediction achieves the highest performance, as evidenced by various metrics, including an average precision (AP) score of 83.39% and an F-1 score of 85.71%. This approach holds promise for developing automated tools to process satellite imagery, benefiting fields such as agriculture, land use monitoring, and disaster response. [ABSTRACT FROM AUTHOR]

Deep learning (DL) techniques have significantly enhanced the potential for wheat head detection in recent times. The different development phases of canopy, genotype, wheat heads, and wheat head orientation provide considerable obstacles. The overlapping density of wheat heads and windinduced picture blurring complicate wheat head recognition. This study describes an effective wheat head detection and counting method. Due to its high throughput in agricultural field analysis, remote sensing phenotyping has grown in popularity. Applying DL methods for image processing and other technological advancements has increased the scope for the quantitative evaluation of various crop traits. The ability to detect and characterize wheat heads in the industry is an important part of the wheat breeding process for selecting high-yielding cultivars. The proposed method uses the Mask region-based convolutional neural network (RCNN) framework to detect and classify the wheat ears. The complete detection task is done in three steps: region proposal generation, region of interest alignment, and mask generation. The global wheat head detection (GWHD) dataset is used for the experimental analysis of the dataset. The proposed method achieved an accuracy of 95.11% on the GWHD dataset, demonstrating its effectiveness in wheat head detection and classification tasks. [ABSTRACT FROM AUTHOR]

Pneumonia is a condition characterised by the sudden inflammation of lung tissue, which is triggered by microorganisms such as fungi, viruses, and bacteria. Chest X-ray imaging (CXR) can detect pneumonia, but it requires considerable time and medical expertise. Consequently, the objective of this study is to diagnose pneumonia using CXR imaging in order to effectively detect early cases of pneumonitis in children. The study employs the ensemble transfer learning convolutional neural network (ETL-CNN) transfer learning ensemble, which combines multiple CNN transfer learning models. Resnet50-VGG19 and VGG19-Xception are the ETL-CNN models used in this investigation. Comparing ETL-CNN models to CNN transfer learning models such as Resnet50, VGG19, and Xception. Pediatric CXR pneumonia, which consists of a normal and pneumonia image, is the source of these study results. The results of this analysis indicate that Resnet50-VGG19 achieved the highest level of accuracy, 99.14%. Additionally, the Resnet50-VGG19 obtained the highest levels of precision and recall when comparing to other models. Consequently, the conclusion of this study is that the Resnet50-VGG19 model can generate acceptable classification performance for pediatric pneumonia based on CXR. This study improves classification results for performance when compared to earlier studies. [ABSTRACT FROM AUTHOR]

Food image recognition is essential for developing an elderly-friendly daily food recording application in Indonesia. However, existing datasets and models are limited and do not cover the diversity and complexity of Indonesian food. In this paper, we present a new dataset of 24,427 images of 160 types of Indonesian food with higher variety and quality than previous datasets. We also train and compare the performance of 67 models based on 16 state-of-the-art deep learning architectures on this dataset. We find that efficientnet_v2_l provides the best accuracy of 85.44%, followed by other models such as convnext_large and swin_s. We also discuss the trade-off between model size and performance, as well as the challenges and limitations of food image classification. Our dataset and models can serve as a basis for developing a user-friendly and accurate food recording application for the elderly population in Indonesia. [ABSTRACT FROM AUTHOR]

Glaucoma is a significant eye disease that can lead to irreversible vision loss if not detected and treated early. This research focuses on developing an automated glaucoma detection system using a combination of a convolutional neural network (CNN) with the residual network 18 (ResNet18) architecture, locality sensitive hashing (LSH), and Hamming distance calculation. The CNN model is trained to extract meaningful features from retinal images, while LSH enables efficient indexing and retrieval of similar images. Hamming distance calculations are utilized to measure the dissimilarity between binary codes obtained from LSH. A dataset of 506 retinal images, consisting of 117 glaucoma images, 19 glaucoma suspect images, and 370 healthy images. The proposed glaucoma detection system achieved an average accuracy of 99.96%, sensitivity of 99.97%, and specificity of 99.94% during training, and 82.37% accuracy, 86.78% sensitivity, and 73.55% specificity during testing. Comparative analysis demonstrated its superiority over traditional methods. Further research should focus on larger datasets and explore multi-class classification for different glaucoma stages. The proposed system has potential for early glaucoma detection, facilitating timely intervention, and preventing vision loss. [ABSTRACT FROM AUTHOR]

There are instances where image compression becomes necessary; however, the use of lossy compression techniques often results in visual artifacts. These artifacts typically remove high-frequency detail and may introduce noise or small image structures. To mitigate the impact of compression on image perception, various technologies, including machine learning and optimization metaheuristics that optimize the parameters of image compression algorithms, have been developed. This paper investigates the application of convolutional neural networks (CNNs) to reduce artifacts associated with image compression, and it presents a proposed method termed deblurring compression image using a CNN (DCI-CNN). Trained on a UTKFace dataset and tested on six benchmark images, the DCI-CNN aims to address artifacts such as block artifacts, ringing artifacts, blurring artifacts, color bleeding, and mosquito noise. The DCI-CNN application is designed to enhance the visual quality and fidelity of compressed images, offering a more detailed output compared to generic and other deep learningbased deblurring methods found in related work. [ABSTRACT FROM AUTHOR]

Rootstocks are gaining importance in viticulture as a strategy to combat abiotic challenges, as well as enhancing scion physiology. Photosynthetic parameters such as maximum rate of carboxylation of RuBP (Vcmax) and the maximum rate of electron transport driving RuBP regeneration (Jmax) have been identified as ideal targets for potential influence by rootstock and breeding. However, leaf specific direct measurement of these photosynthetic parameters is time consuming, limiting the information scope and the number of individuals that can be screened. This study aims to overcome these limitations by employing hyperspectral imaging combined with artificial intelligence (AI) to predict these key photosynthetic traits at the canopy level. Hyperspectral imaging captures detailed optical properties across a broad range of wavelengths (400 to 1000 nm), enabling use of all wavelengths in a comprehensive analysis of the entire vine's photosynthetic performance (Vcmax and Jmax). Artificial intelligence-based prediction models that blend the strength of deep learning and machine learning were developed using two growing seasons data measured post-solstice at 15 h, 14 h, 13 h and 12 h daylengths for Vitis hybrid 'Marquette' grafted to five commercial rootstocks and 'Marquette' grafted to 'Marquette'. Significant differences in photosynthetic efficiency (Vcmax and Jmax) were noted for both direct and indirect measurements for the six rootstocks, indicating that rootstock genotype and daylength have a significant influence on scion photosynthesis. Evaluation of multiple feature-extraction algorithms indicated the proposed Vitis base model incorporating a 1D-Convolutional neural Network (CNN) had the best prediction performance with a R2 of 0.60 for Vcmax and Jmax. Inclusion of weather and chlorophyll parameters slightly improved model performance for both photosynthetic parameters. Integrating AI with hyperspectral remote phenotyping provides potential for high-throughput whole vine assessment of photosynthetic performance and selection of rootstock genotypes that confer improved photosynthetic performance potential in the scion. [ABSTRACT FROM AUTHOR]

Introduction: Malignant pleural mesothelioma (MPM) is a poor-prognosis disease. Owing to the recent availability of new therapeutic options, there is a need to better assess prognosis. The initial clinical response could represent a useful parameter. Methods: We proposed a transfer learning approach to predict an initial treatment response starting from baseline CT scans of patients with advanced/ unresectable MPM undergoing first-line systemic therapy. The therapeutic response has been assessed according to the mRECIST criteria by CT scan at baseline and after two to three treatment cycles. We used three slices of baseline CT scan as input to the pre-trained convolutional neural network as a radiomic feature extractor. We identified a feature subset through a double feature selection procedure to train a binary SVM classifier to discriminate responders (partial response) from non-responders (stable or disease progression). Results: The performance of the prediction classifiers was evaluated with an 80:20 hold-out validation scheme. We have evaluated how the developed model was robust to variations in the slices selected by the radiologist. In our dataset, 25 patients showed an initial partial response, whereas 13 patients showed progressive or stable disease. On the independent test, the proposed model achieved a median AUC and accuracy of 86.67% and 87.50%, respectively. Conclusions: The proposed model has shown high performance even by varying the reference slices. Novel tools could help to improve the prognostic assessment of patients with MPM and to better identify subgroups of patients with different therapeutic responsiveness. [ABSTRACT FROM AUTHOR]

Fabric defects in the conventional manufacturing of acoustic panels are detected via manual visual inspections, which are prone to problems due to human errors. Implementing an automated fabric inspection system can improve productivity and increase product quality. In this work, advanced machine learning (ML) techniques for fabric defect detection are reviewed, and two deep learning (DL) models are developed using transfer learning based on pre-trained convolutional neural network (CNN) architectures. The dataset used for this work consists of 1800 images with six different classes, made up of one class of fabric in good condition and five classes of fabric defects. The model design process involves pre-processing of the images, modification of the neural network layers, as well as selection and optimisation of the network's hyperparameters. The average accuracies of the two CNN models developed in this work, which used the GoogLeNet and the ResNet50 architectures, are 89.84% and 95.45%, respectively, showing statistically significant results. The interpretability of the models is discussed using the Grad-CAM technique. Relevant image acquisition hardware requirements are also put forward for integration with the detection software, which can enable successful deployment of the model for the automated fabric inspection. [ABSTRACT FROM AUTHOR]

The construction industry plays an important role in China's economy. However, traditional construction quality inspection methods have problems such as low efficiency, high leakage rate, and poor real-time performance. Therefore, taking the appearance quality defects of reinforced concrete engineering as an example, this study combines building information models and computer vision technology. By this way, it can achieve automatic quality inspection and intelligent quality problem prevention, to improve the quality management level during the construction process. In this study, ResNet50 network was pre-trained and classified using a self-made defect image database. After transfer learning, the accuracy values of the training and testing sets remained stable at around 0.95 and the loss value remained stable at around 0.10 after 10 epochs, indicating a significant improvement in learning performance. In the defect area quantification experiment, four corner coordinates of the defect image were calculated. These corner coordinates are simultaneously obtained and saved during on-site image acquisition. According to the calculation results, the defect area of honeycomb is approximately 67.67 square centimetres. These results confirm that this method improves the efficiency and accuracy of quality inspection and has potential application in quality inspection of prefabricated building construction. [ABSTRACT FROM AUTHOR]

The rapid growth of Internet of Things (IoT) devices has brought numerous benefits to the interconnected world. However, the ubiquitous nature of IoT networks exposes them to various security threats, including anomaly intrusion attacks. In addition, IoT devices generate a high volume of unstructured data. Traditional intrusion detection systems often struggle to cope with the unique characteristics of IoT networks, such as resource constraints and heterogeneous data sources. Given the unpredictable nature of network technologies and diverse intrusion methods, conventional machine-learning approaches seem to lack efficiency. Across numerous research domains, deep learning techniques have demonstrated their capability to precisely detect anomalies. This study designs and enhances a novel anomaly-based intrusion detection system (AIDS) for IoT networks. Firstly, a Sparse Autoencoder (SAE) is applied to reduce the high dimension and get a significant data representation by calculating the reconstructed error. Secondly, the Convolutional Neural Network (CNN) technique is employed to create a binary classification approach. The proposed SAE-CNN approach is validated using the Bot-IoT dataset. The proposed models exceed the performance of the existing deep learning approach in the literature with an accuracy of 99.9%, precision of 99.9%, recall of 100%, F1 of 99.9%, False Positive Rate (FPR) of 0.0003, and True Positive Rate (TPR) of 0.9992. In addition, alternative metrics, such as training and testing durations, indicated that SAE-CNN performs better. [ABSTRACT FROM AUTHOR]

The aftermath of a natural disaster leaves victims trapped in rubble which is challenging to detect by smart drones due to the victims in low visibility under the adverse disaster environments and victims in various sizes. To overcome the above challenges, a transformer fusion-based scale-aware attention network (TFSANet) is proposed to overcome adverse environmental impacts in disaster areas by robustly integrating the latent interactions between RGB and thermal images and to address the problem of various-sized victim detection. Firstly, a transformer fusion model is developed to incorporate a two-stream backbone network to effectively fuse the complementary characteristics between RGB and thermal images. This aims to solve the problem that the victims cannot be seen clearly due to the adverse disaster area, such as smog and heavy rain. In addition, a scale-aware attention mechanism is designed to be embedded into the head network to adaptively adjust the size of receptive fields aiming to capture victims with different scales. Extensive experiments on two challenging datasets indicate that our TFSANet achieves superior results. The proposed method achieves 86.56% average precision (AP) on the National Institute of Informatics—Chiba University (NII-CU) multispectral aerial person detection dataset, outperforming the state-of-the-art approach by 4.38%. On the drone-captured RGBT person detection (RGBTDronePerson) dataset, the proposed method significantly improves the AP of the state-of-the-art approach by 4.33%. [ABSTRACT FROM AUTHOR]