Search

In the context of TinyML, many research efforts have been devoted to designing forward topologies to support On-Device Learning. Reaching this target would bring numerous advantages, including reductions in latency and computational complexity, stronger privacy, data safety and robustness to adversarial attacks, higher resilience against concept drift, etc. However, On-Device Learning on resource constrained devices poses severe limitations to computational power and memory. Therefore, deploying Neural Networks on tiny devices appears to be prohibitive, since their backpropagation-based training is too memory demanding for their embedded assets. Using Extreme Learning Machines based on Convolutional Neural Networks might be feasible and very convenient, especially for Feature Extraction tasks. However, it requires searching for a randomly initialized topology that achieves results as good as those achieved by the backpropagated model. This work proposes a novel approach for automatically composing an Extreme Convolutional Feature Extractor, based on Neural Architecture Search and Bayesian Optimization. It was applied to the CIFAR-10 and MNIST datasets for evaluation. Two search spaces have been defined, as well as a search strategy that has been tested with two surrogate models, Gaussian Process and Random Forest. A performance estimation strategy was defined, keeping the feature set computed by the MLCommons-Tiny benchmark ResNet as a reference model. In as few as 1200 search iterations, the proposed strategy was able to achieve a topology whose extracted features scored a mean square error equal to 0.64 compared to the reference set. Further improvements are required, with a target of at least one order of magnitude decrease in mean square error for improved classification accuracy. The code is made available via GitHub to allow for the reproducibility of the results reported in this paper. [ABSTRACT FROM AUTHOR]

Algorithms optimized for high-performance computing, which ensure both speed and accuracy, are crucial for real-time data analysis in heavy-ion physics experiments. The application of neural networks and other machine learning methodologies, which are fast and have high accuracy, in physics experiments has become increasingly popular over recent years. This paper introduces a fast neural network package named ANN4FLES developed in C++, which has been optimized for use on a high-performance computing cluster for the future Compressed Baryonic Matter (CBM) experiment at the Facility for Antiproton and Ion Research (FAIR, Darmstadt, Germany). The use of neural networks for classifying events during heavy-ion collisions in the CBM experiment is under investigation. This paper provides a detailed description of the application of ANN4FLES in identifying collisions where a quark–gluon plasma (QGP) was produced. The methodology detailed here will be used in the development of a QGP trigger for event selection within the First Level Event Selection (FLES) package for the CBM experiment. Fully-connected and convolutional neural networks have been created for the identification of events containing QGP, which are simulated with the Parton–Hadron–String Dynamics (PHSD) microscopic off-shell transport approach, for central Au + Au collisions at an energy of 31.2 A GeV. The results show that the convolutional neural network outperforms the fully-connected networks and achieves over 95% accuracy on the testing dataset. [ABSTRACT FROM AUTHOR]

The COVID-19 pandemic has posed significant challenges in accurately diagnosing the disease, as severe cases may present symptoms similar to pneumonia. Real-Time Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) is the conventional diagnostic technique; however, it has limitations in terms of time-consuming laboratory procedures and kit availability. Radiological chest images, such as X-rays and Computed Tomography (CT) scans, have been essential in aiding the diagnosis process. In this research paper, we propose a deep learning (DL) approach based on Convolutional Neural Networks (CNNs) to enhance the detection of COVID-19 and its variants from chest X-ray images. Building upon the existing research in SARS and COVID-19 identification using AI and machine learning techniques, our DL model aims to extract the most significant features from the X-ray scans of affected individuals. By employing an explanatory CNN-based technique, we achieved a promising accuracy of up to 97 % in detecting COVID-19 cases, which can assist physicians in effectively screening and identifying probable COVID-19 patients. This study highlights the potential of DL in medical imaging, specifically in detecting COVID-19 from radiological images. The improved accuracy of our model demonstrates its efficacy in aiding healthcare professionals and mitigating the spread of the disease. [ABSTRACT FROM AUTHOR]

Due to the need to know the availability of solar resources for the solar renewable technologies in advance, this paper presents a new methodology based on computer vision and the object detection technique that uses convolutional neural networks (EfficientDet-D2 model) to detect clouds in image series. This methodology also calculates the speed and direction of cloud motion, which allows the prediction of transients in the available solar radiation due to clouds. The convolutional neural network model retraining and validation process finished successfully, which gave accurate cloud detection results in the test. Also, during the test, the estimation of the remaining time for a transient due to a cloud was accurate, mainly due to the precise cloud detection and the accuracy of the remaining time algorithm. [ABSTRACT FROM AUTHOR]

This document is a special issue of the journal "Algorithms" focused on algorithms in data classification. It provides an overview of different subcategories of data classification, such as binary classification, multi-class classification, and text classification. The issue includes twelve papers that cover a wide range of applications, including class imbalance in Gaussian mixture models, blood cell classification using deep learning techniques, student dropout prediction in online education, and energy consumption in industrial plants. Other topics covered include the use of artificial intelligence models for rehabilitation guidance, hyperparameter optimization of artificial neural networks, feature selection in big data classification, and the classification of acute psychological stress and physical activity using wristband devices. The issue also includes specialized software for data classification and the use of artificial intelligence models for intelligent wear monitoring. [Extracted from the article]

Timely information about the need to thin forests is vital in forest management to maintain a healthy forest while maximizing income. Currently, very-high-spatial-resolution remote sensing data can provide crucial assistance to experts when evaluating the maturity of thinnings. Nevertheless, this task is still predominantly carried out in the field and demands extensive resources. This paper presents a deep convolutional neural network (DCNN) to detect the necessity and urgency of carrying out thinnings using only remote sensing data. The approach uses very-high-spatial-resolution RGB and near-infrared orthophotos; a canopy height model (CHM); a digital terrain model (DTM); the slope; and reference data, which, in this case, originate from spruce-dominated forests in the Austrian Alps. After tuning, the model achieves an F1 score of 82.23% on our test data, which indicates that the model is usable in a practical setting. We conclude that DCNNs are capable of detecting the need to carry out thinnings in forests. In contrast, attempts to assess the urgency of the need for thinnings with DCNNs proved to be unsuccessful. However, additional data, such as age or yield class, have the potential to improve the results. Our investigation into the influence of each individual input feature shows that orthophotos appear to contain the most relevant information for detecting the need for thinning. Moreover, we observe a gain in performance when adding the CHM and slope, whereas adding the DTM harms the model's performance. [ABSTRACT FROM AUTHOR]

Currently, video and digital images possess extensive utility, ranging from recreational and social media purposes to verification, military operations, legal proceedings, and penalization. The enhancement mechanisms of this medium have undergone significant advancements, rendering them more accessible and widely available to a larger population. Consequently, this has facilitated the ease with which counterfeiters can manipulate images. Convolutional neural network (CNN)-based feature extraction and detection techniques were used to carry out this task, which aims to identify the variations in image features between modified and non-manipulated areas. However, the effectiveness of the existing detection methods could be more efficient. The contributions of this paper include the introduction of a segmentation method to identify the forgery region in images with the U-Net model's improved structure. The suggested model connects the encoder and decoder pipeline by improving the convolution module and increasing the set of weights in the U-Net contraction and expansion path. In addition, the parameters of the U-Net network are optimized by using the grasshopper optimization algorithm (GOA). Experiments were carried out on the publicly accessible image tempering detection evaluation dataset from the Chinese Academy of Sciences Institute of Automation (CASIA) to assess the efficacy of the suggested strategy. The results show that the U-Net modifications significantly improve the overall segmentation results compared to other models. The effectiveness of this method was evaluated on CASIA, and the quantitative results obtained based on accuracy, precision, recall, and the F1 score demonstrate the superiority of the U-Net modifications over other models. [ABSTRACT FROM AUTHOR]

Alzheimer's disease is a chronic neurodegenerative disease that causes brain cells to degenerate, resulting in decreased physical and mental abilities and, in severe cases, permanent memory loss. It is considered as the most common and fatal form of dementia. Although mild cognitive impairment (MCI) precedes Alzheimer's disease (AD), it does not necessarily show the obvious symptoms of AD. As a result, it becomes challenging to distinguish between mild cognitive impairment and cognitively normal. In this paper, we propose an ensemble of deep learners based on convolutional neural networks for the early diagnosis of Alzheimer's disease. The proposed approach utilises simple averaging ensemble and weighted averaging ensemble methods. The ensemble-based transfer learning model demonstrates enhanced generalization and performance for AD diagnosis compared to traditional transfer learning methods. Extensive experiments on the OASIS-3 dataset validate the effectiveness of the proposed model, showcasing its superiority over state-of-the-art transfer learning approaches in terms of accuracy, robustness, and efficiency. [ABSTRACT FROM AUTHOR]

The recognition of human activities using vision-based techniques has become a crucial research field in video analytics. Over the last decade, there have been numerous advancements in deep learning algorithms aimed at accurately detecting complex human actions in video streams. While these algorithms have demonstrated impressive performance in activity recognition, they often exhibit a bias towards either model performance or computational efficiency. This biased trade-off between robustness and efficiency poses challenges when addressing complex human activity recognition problems. To address this issue, this paper presents a computationally efficient yet robust approach, exploiting saliency-aware spatial and temporal features for human action recognition in videos. To achieve effective representation of human actions, we propose an efficient approach called the dual-attentional Residual 3D Convolutional Neural Network (DA-R3DCNN). Our proposed method utilizes a unified channel-spatial attention mechanism, allowing it to efficiently extract significant human-centric features from video frames. By combining dual channel-spatial attention layers with residual 3D convolution layers, the network becomes more discerning in capturing spatial receptive fields containing objects within the feature maps. To assess the effectiveness and robustness of our proposed method, we have conducted extensive experiments on four well-established benchmark datasets for human action recognition. The quantitative results obtained validate the efficiency of our method, showcasing significant improvements in accuracy of up to 11% as compared to state-of-the-art human action recognition methods. Additionally, our evaluation of inference time reveals that the proposed method achieves up to a 74× improvement in frames per second (FPS) compared to existing approaches, thus showing the suitability and effectiveness of the proposed DA-R3DCNN for real-time human activity recognition. [ABSTRACT FROM AUTHOR]

Channel identification is a useful function to support wireless telecommunication operations because the knowledge of the radio frequency propagation channel characteristics can improve communication efficiency and robustness. In recent times, the application of machine learning (ML) algorithms to the problem of channel identification has been proposed in the literature. In particular, Deep Learning (DL) has demonstrated superior performance to 'shallow' machine learning algorithms for many wireless communication functions. Inspired by the success of DL in literature, the authors in this paper apply Convolutional Neural Networks (CNN) to the problem of channel identification, which is still an emerging research area. CNN is a deep learning algorithm that has demonstrated superior performance to ML algorithms, in particular for image processing tasks. Because the digitized RF signal is a one-dimensional time series, different algorithms are applied to convert the time series to images using various Time Frequency Transform (TFT) including the CWTs, spectrogram, and Wigner Ville distribution. The images are then provided as input to the CNN. The approach is applied to a data set based on weather radar pulse signals generated in the laboratory of the author's facilities on which different fading models are applied. These models are inspired by the tap-delay-line 3GPP configurations defined in the standards, but they have been customized with Nakagami-m fading distribution (3GPP-like fading models). The results show the superior performance of time–frequency CNN in comparison to 1D CNN for different values of Signal to Noise Ratio (SNR) in dB. In particular, the study shows that the Continuous Wavelet Transform (CWT) has the optimal performance in this data set, but the choice of the mother wavelet remains a problem to be solved (this is a well-known problem in the research literature). Then, this study also proposes an adaptive technique for the choice of the optimal mother wavelet, which is evaluated on the mentioned data set. The results show that the adaptive proposed approach is able to obtain the optimal performance for most of the SNR conditions. [ABSTRACT FROM AUTHOR]

Residential and industrial buildings are significant consumers of energy, which can be reduced by controlling their respective Heating, Ventilation, and Air Conditioning (HVAC) systems. Demand-based Ventilation (DCV) determines the operational times of ventilation systems that depend on indoor air quality (IAQ) conditions, including CO2 concentration changes, and the occupants' comfort requirements. The prediction of CO2 concentration changes can act as a proxy estimator of occupancy changes and provide feedback about the utility of current ventilation controls. This paper proposes a Hierarchical Model for CO2 Variation Predictions (HMCOVP) to accurately predict these variations. The proposed framework addresses two concerns in state-of-the-art implementations. First, the hierarchical structure enables fine-tuning of the produced models, facilitating their transferability to different spatial settings. Second, the formulation incorporates time dependencies, defining the relationship between different IAQ factors. Toward that goal, the HMCOVP decouples the variation prediction into two complementary steps. The first step transforms lagged versions of environmental features into image representations to predict the variations' direction. The second step combines the first step's result with environment-specific historical data to predict CO2 variations. Through the HMCOVP, these predictions, which outperformed state-of-the-art approaches, help the ventilation systems in their decision-making processes, reducing energy consumption and carbon-based emissions. [ABSTRACT FROM AUTHOR]

The field of application of data-driven product development is diverse and ranges from requirements through the early phases to the detailed design of the product. The goal is to consistently analyze data to support and improve individual steps in the development process. In the context of this work, the focus is on the design and detailing phase, represented by the virtual testing of products through Finite Element (FE) simulations. However, due to the heterogeneous data of a simulation model, automatic use is a big challenge. A method is therefore presented that utilizes the entire stock of calculated simulations to predict the plausibility of new simulations. Correspondingly, a large amount of data is utilized to support less experienced users of FE software in the application. Thus, obvious errors in the simulation should be detected immediately with this procedure and unnecessary iterations are therefore avoided. Previous solutions were only able to perform a general plausibility classification, whereas the approach presented in this paper is intended to predict specific error sources in FE simulations. [ABSTRACT FROM AUTHOR]

Creating machines that behave and work in a way similar to humans is the objective of artificial intelligence (AI). In addition to pattern recognition, planning, and problem-solving, computer activities with artificial intelligence include other activities. A group of algorithms called "deep learning" is used in machine learning. With the aid of magnetic resonance imaging (MRI), deep learning is utilized to create models for the detection and categorization of brain tumors. This allows for the quick and simple identification of brain tumors. Brain disorders are mostly the result of aberrant brain cell proliferation, which can harm the structure of the brain and ultimately result in malignant brain cancer. The early identification of brain tumors and the subsequent appropriate treatment may lower the death rate. In this study, we suggest a convolutional neural network (CNN) architecture for the efficient identification of brain tumors using MR images. This paper also discusses various models such as ResNet-50, VGG16, and Inception V3 and conducts a comparison between the proposed architecture and these models. To analyze the performance of the models, we considered different metrics such as the accuracy, recall, loss, and area under the curve (AUC). As a result of analyzing different models with our proposed model using these metrics, we concluded that the proposed model performed better than the others. Using a dataset of 3264 MR images, we found that the CNN model had an accuracy of 93.3%, an AUC of 98.43%, a recall of 91.19%, and a loss of 0.25. We may infer that the proposed model is reliable for the early detection of a variety of brain tumors after comparing it to the other models. [ABSTRACT FROM AUTHOR]

3D mesh as a complex data structure can provide effective shape representation for 3D objects, but due to the irregularity and disorder of the mesh data, it is difficult for convolutional neural networks to be directly applied to 3D mesh data processing. At the same time, the extensive use of convolutional kernels and pooling layers focusing on local features can cause the loss of spatial information and dependencies of low-level features. In this paper, we propose a self-attentive convolutional network MixFormer applied to 3D mesh models. By defining 3D convolutional kernels and vector self-attention mechanisms applicable to 3D mesh models, our neural network is able to learn 3D mesh model features. Combining the features of convolutional networks and transformer networks, the network can focus on both local detail features and long-range dependencies between features, thus achieving good learning results without stacking multiple layers and saving arithmetic overhead compared to pure transformer architectures. We conduct classification and semantic segmentation experiments on SHREC15, SCAPE, FAUST, MIT, and Adobe Fuse datasets. Experimental results show that the network can achieve 96.7% classification and better segmentation results by using fewer parameters and network layers. [ABSTRACT FROM AUTHOR]

Digitization and transcription of historic documents offer new research opportunities for humanists and are the topics of many edition projects. However, manual work is still required for the main phases of layout recognition and the subsequent optical character recognition (OCR) of early printed documents. This paper describes and evaluates how deep learning approaches recognize text lines and can be extended to layout recognition using background knowledge. The evaluation was performed on five corpora of early prints from the 15th and 16th Centuries, representing a variety of layout features. While the main text with standard layouts could be recognized in the correct reading order with a precision and recall of up to 99.9%, also complex layouts were recognized at a rate as high as 90% by using background knowledge, the full potential of which was revealed if many pages of the same source were transcribed. [ABSTRACT FROM AUTHOR]

Over the last few years, human activity recognition (HAR) has drawn increasing interest from the scientific community. This attention is mainly attributable to the proliferation of wearable sensors and the expanding role of HAR in such fields as healthcare, sports, and human activity monitoring. Convolutional neural networks (CNN) are becoming a popular approach for addressing HAR problems. However, this method requires extensive training datasets to perform adequately on new data. This paper proposes a novel deep learning model pre-trained on scalograms generated using the continuous wavelet transform (CWT). Nine popular CNN architectures and different CWT configurations were considered to select the best performing combination, resulting in the training and evaluation of more than 300 deep learning models. On the source KU-HAR dataset, the selected model achieved classification accuracy and an F1 score of 97.48% and 97.52%, respectively, which outperformed contemporary state-of-the-art works where this dataset was employed. On the target UCI-HAPT dataset, the proposed model resulted in a maximum accuracy and F1-score increase of 0.21% and 0.33%, respectively, on the whole UCI-HAPT dataset and of 2.82% and 2.89%, respectively, on the UCI-HAPT subset. It was concluded that the usage of the proposed model, particularly with frozen layers, results in improved performance, faster training, and smoother gradient descent on small HAR datasets. However, the use of the pre-trained model on sufficiently large datasets may lead to negative transfer and accuracy degradation. [ABSTRACT FROM AUTHOR]

The objective of this work is to show the application of a Deep Learning algorithm able to operate the segmentation of ancient Egyptian hieroglyphs present in an image, with the ambition to be as versatile as possible despite the variability of the image source. The problem is quite complex, the main obstacles being the considerable amount of different classes of existing hieroglyphs, the differences related to the hand of the scribe as well as the great differences among the various supports, such as papyri, stone or wood, where they are written. Furthermore, as in all archaeological finds, damage to the supports are frequent, with the consequence that hieroglyphs can be partially corrupted. In order to face this challenging problem, we leverage on the well-known Detectron2 platform, developed by the Facebook AI Research Group, focusing on the Mask R-CNN architecture to perform segmentation of image instances. Likewise, for several machine learning studies, one of the hardest challenges is the creation of a suitable dataset. In this paper, we will describe a hieroglyph dataset that has been created for the purpose of segmentation, highlighting its pros and cons, and the impact of different hyperparameters on the final results. Tests on the segmentation of images taken from public databases will also be presented and discussed along with the limitations of our study. [ABSTRACT FROM AUTHOR]

Machine Learning (ML) and/or Deep Learning (DL) methods can be used to predict fluid flow in porous media, as a suitable replacement for classical numerical approaches. Such data-driven approaches attempt to learn mappings between finite-dimensional Euclidean spaces. A novel neural framework, named Fourier Neural Operator (FNO), has been recently developed to act on infinite-dimensional spaces. A high proportion of the research available on the FNO has focused on problems with large-shape data. Furthermore, most published studies apply the FNO method to existing datasets. This paper applies and evaluates FNO to predict pressure distribution over a small, specified shape-data problem using 1700 Finite Element Method (FEM) generated samples, from heterogeneous permeability fields as the input. Considering FEM-calculated outputs as the true values, the configured FNO model provides superior prediction performance to that of a Convolutional Neural Network (CNN) in terms of statistical error assessment based on the coefficient of determination (R 2 ) and Mean Squared Error (MSE). Sensitivity analysis considering a range of FNO configurations reveals that the most accurate model is obtained using m o d e s = 15 and w i d t h = 100 . Graphically, the FNO model precisely follows the observed trend in each porous medium evaluated. There is potential to further improve the FNO's performance by including physics constraints in its network configuration. [ABSTRACT FROM AUTHOR]

Atrial fibrillation (AF) is an arrhythmia that may cause blood clots and increase the risk of stroke and heart failure. Traditional 12-lead electrocardiogram (ECG) acquisition equipment is complex and difficult to carry. Short single-lead ECG recordings based on wearable devices can remedy these shortcomings. However, reliable and accurate atrial fibrillation detection is still an issue because of the limited information on the short single-lead ECG recordings. In this paper, we propose a novel multi-branch convolutional neural network and bidirectional long short-term memory network (MCNN-BLSTM) to deal with the reliability and accuracy of AF detection in short single-lead ECG recordings. Firstly, to fuller extract the feature information of short single-lead ECG recordings, the MCNN module is designed to dynamically set several corresponding branches according to the number of slices of short single-lead ECG recordings. Then, the BLSTM module is designed to further enhance the feature information learned from each branch. We validated the model on the PhysioNet/CinC Challenge 2017 (CinC2017) database and verified the generalization of the model on the China Physiological Signal Challenge 2018 (CPSC2018) database. The results show that the accuracy of the model on the CinC 2017 database reaches 87.57%, and the average F1 score reaches 84.56%. The accuracy of the model on the CPSC 2018 database reaches 87.50%, and the average F1 score reaches 82.01%. Compared with other advanced methods, our model shows better performance and can meet the daily needs of atrial fibrillation detection with short ECG wearable devices. [ABSTRACT FROM AUTHOR]

This paper focuses on the gaps that occur inside plantations; these gaps, although not having anything growing in them, still happen to be watered. This action ends up wasting tons of liters of water every year, which translates into financial and environmental losses. To avoid these losses, we suggest early detection. To this end, we analyzed the different available neural networks available with multispectral images. This entailed training each regional and regression-based network five times with five different datasets. Networks based on two possible solutions were chosen: unmanned aerial vehicle (UAV) depletion or post-processing with external software. The results show that the best network for UAV depletion is the Tiny-YOLO (You Only Look Once) version 4-type network, and the best starting weights for Mask-RCNN were from the Tiny-YOLO network version. Although no mean average precision (mAP) of over 70% was achieved, the final trained networks managed to detect mostly gaps, including low-vegetation areas and very small gaps, which had a tendency to be overlooked during the labeling stage. [ABSTRACT FROM AUTHOR]

Activity recognition is the process of continuously monitoring a person's activity and movement. Human posture recognition can be utilized to assemble a self-guidance practice framework that permits individuals to accurately learn and rehearse yoga postures without getting help from anyone else. With the use of deep learning algorithms, we propose an approach for the efficient detection and recognition of various yoga poses. The chosen dataset consists of 85 videos with 6 yoga postures performed by 15 participants, where the keypoints of users are extracted using the Mediapipe library. A combination of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) has been employed for yoga pose recognition through real-time monitored videos as a deep learning model. Specifically, the CNN layer is used for the extraction of features from the keypoints and the following LSTM layer understands the occurrence of sequence of frames for predictions to be implemented. In following, the poses are classified as correct or incorrect; if a correct pose is identified, then the system will provide user the corresponding feedback through text/speech. This paper combines machine learning foundations with data structures as the synergy between these two areas can be established in the sense that machine learning techniques and especially deep learning can efficiently recognize data schemas and make them interoperable. [ABSTRACT FROM AUTHOR]

Modern trainable image recognition models are vulnerable to different types of perturbations; hence, the development of resilient intelligent algorithms for safety-critical applications remains a relevant concern to reduce the impact of perturbation on model performance. This paper proposes a model and training method for a resilient image classifier capable of efficiently functioning despite various faults, adversarial attacks, and concept drifts. The proposed model has a multi-section structure with a hierarchy of optimized class prototypes and hyperspherical class boundaries, which provides adaptive computation, perturbation absorption, and graceful degradation. The proposed training method entails the application of a complex loss function assembled from its constituent parts in a particular way depending on the result of perturbation detection and the presence of new labeled and unlabeled data. The training method implements principles of self-knowledge distillation, the compactness maximization of class distribution and the interclass gap, the compression of feature representations, and consistency regularization. Consistency regularization makes it possible to utilize both labeled and unlabeled data to obtain a robust model and implement continuous adaptation. Experiments are performed on the publicly available CIFAR-10 and CIFAR-100 datasets using model backbones based on modules ResBlocks from the ResNet50 architecture and Swin transformer blocks. It is experimentally proven that the proposed prototype-based classifier head is characterized by a higher level of robustness and adaptability in comparison with the dense layer-based classifier head. It is also shown that multi-section structure and self-knowledge distillation feature conserve resources when processing simple samples under normal conditions and increase computational costs to improve the reliability of decisions when exposed to perturbations. [ABSTRACT FROM AUTHOR]

Traditional methods used to identify and monitor insect species are time-consuming, costly, and fully dependent on the observer's ability. This paper presents a deep learning-based cicada species recognition system using acoustic signals to classify the cicada species. The sound recordings of cicada species were collected from different online sources and pre-processed using denoising algorithms. An improved Härmä syllable segmentation method is introduced to segment the audio signals into syllables since the syllables play a key role in identifying the cicada species. After that, a visual representation of the audio signal was obtained using a spectrogram, which was fed to a convolutional neural network (CNN) to perform classification. The experimental results validated the robustness of the proposed method by achieving accuracies ranging from 66.67% to 100%. [ABSTRACT FROM AUTHOR]

This work includes a brief overview of the applications of the powerful and easy-to-perform method of microwave radiometry (MWR) for the diagnosis of various diseases. The main goal of this paper is to develop a method for diagnosing breast oncology based on machine learning algorithms using thermometric data, both real medical measurements and simulation results of MWR examinations. The dataset includes distributions of deep and skin temperatures calculated in numerical models of the dynamics of thermal and radiation fields inside biological tissue. The constructed combined dataset allows us to explore the limits of applicability of the MWR method for detecting weak tumors. We use convolutional neural networks and classic machine learning algorithms (k-nearest neighbors, naive Bayes classifier, support vector machine) to classify data. The construction of Kohonen self-organizing maps to explore the structure of our combined dataset demonstrated differences between the temperatures of patients with positive and negative diagnoses. Our analysis shows that the MWR can detect tumors with a radius of up to 0.5 cm if they are at the stage of rapid growth, when the tumor volume doubling occurs in approximately 100 days or less. The use of convolutional neural networks for MWR provides both high sensitivity ( s e n s = 0.86 ) and specificity ( s p e c = 0.82 ), which is an advantage over other methods for diagnosing breast cancer. A new modified scheme for medical measurements of IR temperature and brightness temperature is proposed for a larger number of points in the breast compared to the classical scheme. This approach can increase the effectiveness and sensitivity of diagnostics by several percent. [ABSTRACT FROM AUTHOR]

This paper deals with the effect of exploiting background knowledge for improving an OMR (Optical Music Recognition) deep learning pipeline for transcribing medieval, monophonic, handwritten music from the 12th–14th century, whose usage has been neglected in the literature. Various types of background knowledge about overlapping notes and text, clefs, graphical connections (neumes) and their implications on the position in staff of the notes were used and evaluated. Moreover, the effect of different encoder/decoder architectures and of different datasets for training a mixed model and for document-specific fine-tuning based on an extended OMR pipeline with an additional post-processing step were evaluated. The use of background models improves all metrics and in particular the melody accuracy rate (mAR), which is based on the insert, delete and replace operations necessary to convert the generated melody into the correct melody. When using a mixed model and evaluating on a different dataset, our best model achieves without fine-tuning and without post-processing a mAR of 90.4%, which is raised by nearly 30% to 93.2% mAR using background knowledge. With additional fine-tuning, the contribution of post-processing is even greater: the basic mAR of 90.5% is raised by more than 50% to 95.8% mAR. [ABSTRACT FROM AUTHOR]

In recent years, the topic of contactless biometric identification has gained considerable traction due to the COVID-19 pandemic. One of the most well-known identification technologies is iris recognition. Determining the classification threshold for large datasets of iris images remains challenging. To solve this issue, it is essential to extract more discriminatory features from iris images. Choosing the appropriate loss function to enhance discrimination power is one of the most significant factors in deep learning networks. This paper proposes a novel iris identification framework that integrates the light-weight MobileNet architecture with customized ArcFace and Triplet loss functions. By combining two loss functions, it is possible to improve the compactness within a class and the discrepancies between classes. To reduce the amount of preprocessing, the normalization step is omitted and segmented iris images are used directly. In contrast to the original SoftMax loss, the EER for the combined loss from ArcFace and Triplet is decreased from 1.11% to 0.45%, and the TPR is increased from 99.77% to 100%. In CASIA-Iris-Thousand, EER decreased from 4.8% to 1.87%, while TPR improved from 97.42% to 99.66%. Experiments have demonstrated that the proposed approach with customized loss using ArcFace and Triplet can significantly improve state-of-the-art and achieve outstanding results. [ABSTRACT FROM AUTHOR]

Recently, some researchers adopted the convolutional neural network (CNN) for time series classification (TSC) and have achieved better performance than most hand-crafted methods in the University of California, Riverside (UCR) archive. The secret to the success of the CNN is weight sharing, which is robust to the global translation of the time series. However, global translation invariance is not the only case considered for TSC. Temporal distortion is another common phenomenon besides global translation in time series. The scale and phase changes due to temporal distortion bring significant challenges to TSC, which is out of the scope of conventional CNNs. In this paper, a CNN architecture with an elastic matching mechanism, which is named Elastic Matching CNN (short for EM-CNN), is proposed to address this challenge. Compared with the conventional CNN, EM-CNN allows local time shifting between the time series and convolutional kernels, and a matching matrix is exploited to learn the nonlinear alignment between time series and convolutional kernels of the CNN. Several EM-CNN models are proposed in this paper based on diverse CNN models. The results for 85 UCR datasets demonstrate that the elastic matching mechanism effectively improves CNN performance. [ABSTRACT FROM AUTHOR]

Capsule networks are a type of neural network that use the spatial relationship between features to classify images. By capturing the poses and relative positions between features, this network is better able to recognize affine transformation and surpass traditional convolutional neural networks (CNNs) when handling translation, rotation, and scaling. The stacked capsule autoencoder (SCAE) is a state-of-the-art capsule network that encodes an image in capsules which each contain poses of features and their correlations. The encoded contents are then input into the downstream classifier to predict the image categories. Existing research has mainly focused on the security of capsule networks with dynamic routing or expectation maximization (EM) routing, while little attention has been given to the security and robustness of SCAEs. In this paper, we propose an evasion attack against SCAEs. After a perturbation is generated based on the output of the object capsules in the model, it is added to an image to reduce the contribution of the object capsules related to the original category of the image so that the perturbed image will be misclassified. We evaluate the attack using an image classification experiment on the Mixed National Institute of Standards and Technology Database (MNIST), Fashion-MNIST, and German Traffic Sign Recognition Benchmark (GTSRB) datasets, and the average attack success rate can reach 98.6%. The experimental results indicate that the attack can achieve high success rates and stealthiness. This finding confirms that the SCAE has a security vulnerability that allows for the generation of adversarial samples. Our work seeks to highlight the threat of this attack and focus attention on SCAE's security. [ABSTRACT FROM AUTHOR]

Analog circuits play an important role in modern electronic systems. Aiming to accurately diagnose the faults of analog circuits, this paper proposes a novel variant of a convolutional neural network, namely, a multi-scale convolutional neural network with a selective kernel (MSCNN-SK). In MSCNN-SK, a multi-scale average difference layer is developed to compute multi-scale average difference sequences, and then these sequences are taken as the input of the model, which enables it to mine potential fault characteristics. In addition, a dynamic convolution kernel selection mechanism is introduced to adaptively adjust the receptive field, so that the feature extraction ability of MSCNN-SK is enhanced. Based on two well-known fault diagnosis circuits, comparison experiments are conducted, and experimental results show that our proposed method achieves higher performance. [ABSTRACT FROM AUTHOR]

The field of explainable artificial intelligence (XAI) aims to build explainable and interpretable machine learning (or deep learning) methods without sacrificing prediction performance. Convolutional neural networks (CNNs) have been successful in making predictions, especially in image classification. These popular and well-documented successes use extremely deep CNNs such as VGG16, DenseNet121, and Xception. However, these well-known deep learning models use tens of millions of parameters based on a large number of pretrained filters that have been repurposed from previous data sets. Among these identified filters, a large portion contain no information yet remain as input features. Thus far, there is no effective method to omit these noisy features from a data set, and their existence negatively impacts prediction performance. In this paper, a novel interaction-based convolutional neural network (ICNN) is introduced that does not make assumptions about the relevance of local information. Instead, a model-free influence score (I-score) is proposed to directly extract the influential information from images to form important variable modules. This innovative technique replaces all pretrained filters found by trial-and-error with explainable, influential, and predictive variable sets (modules) determined by the I-score. In other words, future researchers need not rely on pretrained filters; the suggested algorithm identifies only the variables or pixels with high I-score values that are extremely predictive and important. The proposed method and algorithm were tested on real-world data set and a state-of-the-art prediction performance of 99.8% was achieved without sacrificing the explanatory power of the model. This proposed design can efficiently screen patients infected by COVID-19 before human diagnosis and can be a benchmark for addressing future XAI problems in large-scale data sets. [ABSTRACT FROM AUTHOR]

Arabic text classification is a process to simultaneously categorize the different contextual Arabic contents into a proper category. In this paper, a novel deep learning Arabic text computer-aided recognition (ArCAR) is proposed to represent and recognize Arabic text at the character level. The input Arabic text is quantized in the form of 1D vectors for each Arabic character to represent a 2D array for the ArCAR system. The ArCAR system is validated over 5-fold cross-validation tests for two applications: Arabic text document classification and Arabic sentiment analysis. For document classification, the ArCAR system achieves the best performance using the Alarabiya-balance dataset in terms of overall accuracy, recall, precision, and F1-score by 97.76%, 94.08%, 94.16%, and 94.09%, respectively. Meanwhile, the ArCAR performs well for Arabic sentiment analysis, achieving the best performance using the hotel Arabic reviews dataset (HARD) balance dataset in terms of overall accuracy and F1-score by 93.58% and 93.23%, respectively. The proposed ArCAR seems to provide a practical solution for accurate Arabic text representation, understanding, and classification. [ABSTRACT FROM AUTHOR]

The concentration of PM2.5 is an important index to measure the degree of air pollution. When it exceeds the standard value, it is considered to cause pollution and lower the air quality, which is harmful to human health and can cause a variety of diseases, i.e., asthma, chronic bronchitis, etc. Therefore, the prediction of PM2.5 concentration is helpful to reduce its harm. In this paper, a hybrid model called CNN-BiLSTM-Attention is proposed to predict the PM2.5 concentration over the next two days. First, we select the PM2.5 concentration data in hours from January 2013 to February 2017 of Shunyi District, Beijing. The auxiliary data includes air quality data and meteorological data. We use the sliding window method for preprocessing and dividing the corresponding data into a training set, a validation set, and a test set. Second, CNN-BiLSTM-Attention is composed of the convolutional neural network, bidirectional long short-term memory neural network, and attention mechanism. The parameters of this network structure are determined by the minimum error in the training process, including the size of the convolution kernel, activation function, batch size, dropout rate, learning rate, etc. We determine the feature size of the input and output by evaluating the performance of the model, finding out the best output for the next 48 h. Third, in the experimental part, we use the test set to check the performance of the proposed CNN-BiLSTM-Attention on PM2.5 prediction, which is compared by other comparison models, i.e., lasso regression, ridge regression, XGBOOST, SVR, CNN-LSTM, and CNN-BiLSTM. We conduct short-term prediction (48 h) and long-term prediction (72 h, 96 h, 120 h, 144 h), respectively. The results demonstrate that even the predictions of the next 144 h with CNN-BiLSTM-Attention is better than the predictions of the next 48 h with the comparison models in terms of mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2). [ABSTRACT FROM AUTHOR]

Object detection is a challenging computer vision task with numerous real-world applications. In recent years, the concept of the object relationship model has become helpful for object detection and has been verified and realized in deep learning. Nonetheless, most approaches to modeling object relations are limited to using the anchor-based algorithms; they cannot be directly migrated to the anchor-free frameworks. The reason is that the anchor-free algorithms are used to eliminate the complex design of anchors and predict heatmaps to represent the locations of keypoints of different object categories, without considering the relationship between keypoints. Therefore, to better fuse the information between the heatmap channels, it is important to model the visual relationship between keypoints. In this paper, we present a knowledge-driven network (KDNet)—a new architecture that can aggregate and model keypoint relations to augment object features for detection. Specifically, it processes a set of keypoints simultaneously through interactions between their local and geometric features, thereby allowing the modeling of their relationship. Finally, the updated heatmaps were used to obtain the corners of the objects and determine their positions. The experimental results conducted on the RIDER dataset confirm the effectiveness of the proposed KDNet, which significantly outperformed other state-of-the-art object detection methods. [ABSTRACT FROM AUTHOR]

State-of-the-art semantic segmentation methods rely too much on complicated deep networks and thus cannot train efficiently. This paper introduces a novel Circle-U-Net architecture that exceeds the original U-Net on several standards. The proposed model includes circle connect layers, which is the backbone of ResUNet-a architecture. The model possesses a contracting part with residual bottleneck and circle connect layers that capture context and expanding paths, with sampling layers and merging layers for a pixel-wise localization. The results of the experiment show that the proposed Circle-U-Net achieves an improved accuracy of 5.6676%, 2.1587% IoU (Intersection of union, IoU) and can detect 67% classes greater than U-Net, which is better than current results. [ABSTRACT FROM AUTHOR]

Image segmentation plays an important role in the field of image processing, helping to understand images and recognize objects. However, most existing methods are often unable to effectively explore the spatial information in 3D image segmentation, and they neglect the information from the contours and boundaries of the observed objects. In addition, shape boundaries can help to locate the positions of the observed objects, but most of the existing loss functions neglect the information from the boundaries. To overcome these shortcomings, this paper presents a new cascaded 2.5D fully convolutional networks (FCNs) learning framework to segment 3D medical images. A new boundary loss that incorporates distance, area, and boundary information is also proposed for the cascaded FCNs to learning more boundary and contour features from the 3D medical images. Moreover, an effective post-processing method is developed to further improve the segmentation accuracy. We verified the proposed method on LITS and 3DIRCADb datasets that include the liver and tumors. The experimental results show that the performance of the proposed method is better than existing methods with a Dice Per Case score of 74.5 % for tumor segmentation, indicating the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Image denoising is a challenging research problem that aims to recover noise-free images from those that are contaminated with noise. In this paper, we focus on the denoising of images that are contaminated with additive white Gaussian noise. For this purpose, we propose an ensemble learning model that uses the output of three image denoising models, namely ADNet, IRCNN, and DnCNN, in the ratio of 2:3:6, respectively. The first model (ADNet) consists of Convolutional Neural Networks with attention along with median filter layers after every convolutional layer and a dilation rate of 8. In the case of the second model, it is a feed forward denoising CNN or DnCNN with median filter layers after half of the convolutional layers. For the third model, which is Deep CNN Denoiser Prior or IRCNN, the model contains dilated convolutional layers and median filter layers up to the dilated convolutional layers with a dilation rate of 6. By quantitative analysis, we note that our model performs significantly well when tested on the BSD500 and Set12 datasets. [ABSTRACT FROM AUTHOR]

With the widespread success of deep learning in the two-dimensional field, how to apply deep learning methods from two-dimensional to three-dimensional field has become a current research hotspot. Among them, the polygon mesh structure in the three-dimensional representation as a complex data structure provides an effective shape approximate representation for the three-dimensional object. Although the traditional method can extract the characteristics of the three-dimensional object through the graphical method, it cannot be applied to more complex objects. However, due to the complexity and irregularity of the mesh data, it is difficult to directly apply convolutional neural networks to 3D mesh data processing. Considering this problem, we propose a deep learning method based on a capsule network to effectively classify mesh data. We first design a polynomial convolution template. Through a sliding operation similar to a two-dimensional image convolution window, we directly sample on the grid surface, and use the window sampling surface as the minimum unit of calculation. Because a high-order polynomial can effectively represent a surface, we fit the approximate shape of the surface through the polynomial, use the polynomial parameter as the shape feature of the surface, and add the center point coordinates and normal vector of the surface as the pose feature of the surface. The feature is used as the feature vector of the surface. At the same time, to solve the problem of the introduction of a large number of pooling layers in traditional convolutional neural networks, the capsule network is introduced. For the problem of nonuniform size of the input grid model, the capsule network attitude parameter learning method is improved by sharing the weight of the attitude matrix. The amount of model parameters is reduced, and the training efficiency of the 3D mesh model is further improved. The experiment is compared with the traditional method and the latest two methods on the SHREC15 data set. Compared with the MeshNet and MeshCNN, the average recognition accuracy in the original test set is improved by 3.4% and 2.1%, and the average after fusion of features the accuracy reaches 93.8%. At the same time, under the premise of short training time, this method can also achieve considerable recognition results through experimental verification. The three-dimensional mesh classification method proposed in this paper combines the advantages of graphics and deep learning methods, and effectively improves the classification effect of 3D mesh model. [ABSTRACT FROM AUTHOR]

Solving the person re-identification problem involves making associations between the same person's appearances across disjoint camera views. Further, those associations have to be made on multiple surveillance cameras in order to obtain a more efficient and powerful re-identification system. The re-identification problem becomes particularly challenging in very crowded areas. This mainly happens for two reasons. First, the visibility is reduced and occlusions of people can occur. Further, due to congestion, as the number of possible matches increases, the re-identification is becoming challenging to achieve. Additional challenges consist of variations of lightning, poses, or viewpoints, and the existence of noise and blurring effects. In this paper, we aim to generalize person re-identification by implementing a first attempt of a general system, which is robust in terms of distribution variations. Our method is based on the YOLO (You Only Look Once) model, which represents a general object detection system. The novelty of the proposed re-identification method consists of using a simple detection model, with minimal additional costs, but with results that are comparable with those of the other existing dedicated methods. [ABSTRACT FROM AUTHOR]

Visual traffic surveillance using computer vision techniques can be noninvasive, automated and cost effective. Traffic surveillance systems with the ability to detect, count and classify vehicles can be employed in gathering traffic statistics and achieving better traffic control in intelligent transportation systems. This works well in daylight when the road users are clearly visible to the camera, but it often struggles when the visibility of the scene is impaired by insufficient lighting or bad weather conditions such as rain, snow, haze and fog. Therefore, in this paper, we design a dual input faster region-based convolutional neural network (RCNN) to make full use of the complementary advantages of color and thermal images to detect traffic objects in bad weather. Different from the previous detector, we used halfway fusion to fuse color and thermal images for traffic object detection. Besides, we adopt the polling from multiple layers method to adapt the characteristics of large size differences between objects of traffic targets to accurately identify targets of different sizes. The experimental results show that the present method improves the target recognition accuracy by 7.15% under normal weather conditions and 14.2% under bad weather conditions. This exhibits promising potential for implementation with real-world applications. [ABSTRACT FROM AUTHOR]

In the manuscript, the issue of detecting and segmenting out pavement defects on highway roads is addressed. Specifically, computer vision (CV) methods are developed and applied to the problem based on deep learning of convolutional neural networks (ConvNets). A novel neural network structure is considered, based on a pipeline of three ConvNets and endowed with the capacity for context awareness, which improves grid-based search for defects on orthoframes by considering the surrounding image content—an approach, which essentially draws inspiration from how humans tend to solve the task of image segmentation. Also, methods for assessing the quality of segmentation are discussed. The contribution also describes the complete procedure of working with pavement defects in an industrial setting, involving the workcycle of defect annotation, ConvNet training and validation. The results of ConvNet evaluation provided in the paper hint at a successful implementation of the proposed technique. [ABSTRACT FROM AUTHOR]

In this paper, we explore the effect of using different convolutional layers, batch normalization and the global average pooling layer upon a convolutional neural network (CNN) based gaze tracking system. A novel method is proposed to label the participant's face images as gaze points retrieved from eye tracker while watching videos for building a training dataset that is closer to human visual behavior. The participants can swing their head freely; therefore, the most real and natural images can be obtained without too many restrictions. The labeled data are classified according to the coordinate of gaze and area of interest on the screen. Therefore, varied network architectures are applied to estimate and compare the effects including the number of convolutional layers, batch normalization (BN) and the global average pooling (GAP) layer instead of the fully connected layer. Three schemes, including the single eye image, double eyes image and facial image, with data augmentation are used to feed into neural network to train and evaluate the efficiency. The input image of the eye or face for an eye tracking system is mostly a small-sized image with relatively few features. The results show that BN and GAP are helpful in overcoming the problem to train models and in reducing the amount of network parameters. It is shown that the accuracy is significantly improved when using GAP and BN at the mean time. Overall, the face scheme has a highest accuracy of 0.883 when BN and GAP are used at the mean time. Additionally, comparing to the fully connected layer set to 512 cases, the number of parameters is reduced by less than 50% and the accuracy is improved by about 2%. A detection accuracy comparison of our model with the existing George and Routray methods shows that our proposed method achieves better prediction accuracy of more than 6%. [ABSTRACT FROM AUTHOR]

Deep Learning has dramatically advanced the state of the art in vision, speech and many other areas. Recently, numerous deep learning algorithms have been proposed to solve traditional artificial intelligence problems. In this paper, in order to detect the version that can provide the best trade-off in terms of time and accuracy, convolutional networks of various depths have been implemented. Batch normalization is also considered since it acts as a regularizer and achieves the same accuracy with fewer training steps. For maximizing the yield of the complexity by diminishing, as well as minimizing the loss of accuracy, LSTM neural net layers are utilized in the process. The image sequences are proven to be classified by the LSTM in a more accelerated manner, while managing better precision. Concretely, the more complex the CNN, the higher the percentages of exactitude; in addition, but for the high-rank increase in accuracy, the time was significantly decreased, which eventually rendered the trade-off optimal. The average improvement of performance for all models regarding both datasets used amounted to 42 % . [ABSTRACT FROM AUTHOR]