Search

Apricot has a long history of cultivation and has many varieties and types. The traditional variety identification methods are time-consuming and labor-consuming, posing grand challenges to apricot resource management. Tool development in this regard will help researchers quickly identify variety information. This study photographed apricot fruits outdoors and indoors and constructed a dataset that can precisely classify the fruits using a U-net model (F-score: 99%), which helps to obtain the fruit's size, shape, and color features. Meanwhile, a variety search engine was constructed, which can search and identify variety from the database according to the above features. Besides, a mobile and web application (ApricotView) was developed, and the construction mode can be also applied to other varieties of fruit trees. Additionally, we have collected four difficult-to-identify seed datasets and used the VGG16 model for training, with an accuracy of 97%, which provided an important basis for ApricotView. To address the difficulties in data collection bottlenecking apricot phenomics research, we developed the first apricot database platform of its kind (ApricotDIAP, http://apricotdiap.com/) to accumulate, manage, and publicize scientific data of apricot.

Objectives: Many studies have investigated aberrant functional connectivity (FC) using resting-state functional MRI (rs-fMRI) in subjective tinnitus patients. However, no studies have verified the efficacy of resting-state FC as a diagnostic imaging marker. We established a convolutional neural network (CNN) model based on rs-fMRI FC to distinguish tinnitus patients from healthy controls, providing guidance and fast diagnostic tools for the clinical diagnosis of subjective tinnitus. Methods: A CNN architecture was trained on rs-fMRI data from 100 tinnitus patients and 100 healthy controls using an asymmetric convolutional layer. Additionally, a traditional machine learning model and a transfer learning model were included for comparison with the CNN, and each of the three models was tested on three different brain atlases. Results: Of the three models, the CNN model outperformed the other two models with the highest area under the curve, especially on the Dos_160 atlas (AUC = 0.944). Meanwhile, the model with the best classification performance highlights the crucial role of the default mode network, salience network, and sensorimotor network in distinguishing between normal controls and patients with subjective tinnitus. Conclusion: Our CNN model could appropriately tackle the diagnosis of tinnitus patients using rs-fMRI and confirmed the diagnostic value of FC as measured by rs-fMRI.

The internal mechanisms of dental air turbine handpieces (DATHs) have become increasingly intricate over time. To enhance the operational reliability of dental procedures and guarantee patient safety, this study formulated temporal convolution network (TCN) prediction models with the functions of causality in time sequence, transmitting memory, learning, storing, and fast convergence for monitoring the health and diagnosing the rotor and collet failure of DATHs. A handpiece mimicking a dentist’s hand load of 100 g was employed to repeatedly mill a glass porcelain block back and forth for cutting. An accelerometer was employed to capture vibration signals during free-running of unrestrained operation of the handpiece, aiming to discern the characteristic features of these vibrations. These data were then utilized to create a diagnostic health classification (DHC) for further developing a TCN, a 1D convolutional neural network (CNN), and long short-term memory (LSTM) prediction models. The three frameworks were used and compared for machine learning to establish DHC prediction models for the DATH. The experimental results indicate that, in terms of DHC predicted for the experimental dataset, the square categorical cross-entropy loss function error of the TCN framework was generally lower than that of the 1D CNN, which did not have a memory framework or the drawback of the vanishing gradient problem. In addition, the TCN framework outperformed the LSTM model, which required a longer history to provide sufficient diagnostic ability. Still, high accuracies were achieved both in the direction of feed-drive milling and in the gravity of the handpiece through vibration signals. In general, the failure classification prediction model could accurately predict the health and failure mode of the dental handpiece before the use of the DATH when an embedded sensor was available. Therefore, this model could prove to be a beneficial tool for predicting the deterioration patterns of real dental handpieces in their remaining useful life.

Performance prediction of Spark plays a vital role in cluster resource management and system efficiency improvement. The performance of Spark is affected by several variables, such as the size of the input data, the computational power of the system, and the complexity of the algorithm. At the same time, less research has focused on multi-task performance prediction models for Spark. To address these challenges, we propose a multi-task Spark performance prediction model. The model integrates a multi-head attention mechanism and a convolutional neural network. It implements the prediction of execution times for single or multiple Spark applications. Firstly, the data are dimensionally reduced by a dimensionality reduction algorithm and fed into the model. Secondly, the model integrates a multi-head attention mechanism and a convolutional neural network. It captures complex relationships between data features and uses these features for Spark performance prediction. Finally, we use residual connections to prevent overfitting. To validate the performance of the model, we conducted experiments on four Spark benchmark applications. Compared to the benchmark prediction model, our model obtains better performance metrics. In addition, our model predicts multiple Spark benchmark applications simultaneously and maintains deviations within permissible limits. It provides a novel way for the assessment and optimization of Spark.

Objective: For several years, computer technology has been utilized to diagnose lung nodules. When compared to traditional machine learning methods for image processing, deep-learning methods can improve the accuracy of lung nodule diagnosis by avoiding the laborious pre-processing step of the picture (extraction of fake features, etc.). Our goal is to investigate how well deep-learning approaches classify lung nodule malignancy. Method: We evaluated the performance of deep-learning methods on lung nodule malignancy classification via a systematic literature search. We conducted searches for appropriate articles in the PubMed and ISI Web of Science databases and chose those that employed deep learning to classify or predict lung nodule malignancy for our investigation. The figures were plotted, and the data were extracted using SAS version 9.4 and Microsoft Excel 2010, respectively. Results: Sixteen studies that met the criteria were included in this study. The articles classified or predicted pulmonary nodule malignancy using classification and summarization, using convolutional neural network (CNN), autoencoder (AE), and deep belief network (DBN). The AUC of deep-learning models is typically greater than 90% in articles. It demonstrated that deep learning performed well in the diagnosis and forecasting of lung nodules. Conclusion: It is a thorough analysis of the most recent advancements in lung nodule deep-learning technologies. The advancement of image processing techniques, traditional machine learning techniques, deep-learning techniques, and other techniques have all been applied to the technology for pulmonary nodule diagnosis. Although the deep-learning model has demonstrated distinct advantages in the detection of pulmonary nodules, it also carries significant drawbacks that warrant additional research.

Background and Objective: Hypertension is critical risk factor of fatal cardiovascular diseases and multiple organ damage. Early detection of hypertension even at pre-hypertension stage is helpful in preventing the forthcoming complications. Electrocardiogram (ECG) has been attempted to observe the changes in electrical activities of the hearts of hypertensive patients. To automate the ECG assessment in the detection of hypertension, an interpretable hybrid model is proposed in this paper. Methods: The proposed hybrid framework consists of one dimensional - Convolutional Neural Network architecture with four blocks of convolutional layers, maxpooling followed by dropout layers fused with Support Vector Machine classifier in the final layer. The implemented hybrid model is made explainable and interpretable using Local Interpretable Model-agnostic Explanations (LIME) method. The developed hybrid model is trained and tested for patient-wise classification of ECGs using online Physionet datasets and hospital data. Results: The proposed method achieved highest accuracy of 81.81% in patient-wise ECG classification of online datasets, and highest accuracy of 93.33% in patient-wise ECG classification of hospital datasets as normotensive and hypertensive. The visualization of results showed only one normotensive patient's ECG is misclassified (predicted) as hypertensive, with identification of patient number, among the 15 patients (8 normotensive and 7 hypertensive) ECGs tested. In addition, the LIME results demonstrated an explanation to the predictions of hybrid model by highlighting the features and location of ECG waveform responsible for it, thus making the decision of hybrid model more interpretable. Conclusion: Furthermore, our developed system is implemented as an assisting automated software tool called, HANDI (Hypertensive And Normotensive patient Detection with Interpretability) for real-time validation in clinics for early capture of hypertensive and proper monitoring of the patients.

Performance prediction of Spark plays a vital role in cluster resource management and system efficiency improvement. The performance of Spark is affected by several variables, such as the size of the input data, the computational power of the system, and the complexity of the algorithm. At the same time, less research has focused on multi-task performance prediction models for Spark. To address these challenges, we propose a multi-task Spark performance prediction model. The model integrates a multi-head attention mechanism and a convolutional neural network. It implements the prediction of execution times for single or multiple Spark applications. Firstly, the data are dimensionally reduced by a dimensionality reduction algorithm and fed into the model. Secondly, the model integrates a multi-head attention mechanism and a convolutional neural network. It captures complex relationships between data features and uses these features for Spark performance prediction. Finally, we use residual connections to prevent overfitting. To validate the performance of the model, we conducted experiments on four Spark benchmark applications. Compared to the benchmark prediction model, our model obtains better performance metrics. In addition, our model predicts multiple Spark benchmark applications simultaneously and maintains deviations within permissible limits. It provides a novel way for the assessment and optimization of Spark. [ABSTRACT FROM AUTHOR]

Objective: For several years, computer technology has been utilized to diagnose lung nodules. When compared to traditional machine learning methods for image processing, deep-learning methods can improve the accuracy of lung nodule diagnosis by avoiding the laborious pre-processing step of the picture (extraction of fake features, etc.). Our goal is to investigate how well deep-learning approaches classify lung nodule malignancy. Method: We evaluated the performance of deep-learning methods on lung nodule malignancy classification via a systematic literature search. We conducted searches for appropriate articles in the PubMed and ISI Web of Science databases and chose those that employed deep learning to classify or predict lung nodule malignancy for our investigation. The figures were plotted, and the data were extracted using SAS version 9.4 and Microsoft Excel 2010, respectively. Results: Sixteen studies that met the criteria were included in this study. The articles classified or predicted pulmonary nodule malignancy using classification and summarization, using convolutional neural network (CNN), autoencoder (AE), and deep belief network (DBN). The AUC of deep-learning models is typically greater than 90% in articles. It demonstrated that deep learning performed well in the diagnosis and forecasting of lung nodules. Conclusion: It is a thorough analysis of the most recent advancements in lung nodule deep-learning technologies. The advancement of image processing techniques, traditional machine learning techniques, deep-learning techniques, and other techniques have all been applied to the technology for pulmonary nodule diagnosis. Although the deep-learning model has demonstrated distinct advantages in the detection of pulmonary nodules, it also carries significant drawbacks that warrant additional research. [ABSTRACT FROM AUTHOR]

A network intrusion detection model that fuses a convolutional neural network and a gated recurrent unit is proposed to address the problems associated with the low accuracy of existing intrusion detection models for the multiple classification of intrusions and low accuracy of class imbalance data detection. In this model, a hybrid sampling algorithm combining Adaptive Synthetic Sampling (ADASYN) and Repeated Edited nearest neighbors (RENN) is used for sample processing to solve the problem of positive and negative sample imbalance in the original dataset. The feature selection is carried out by combining Random Forest algorithm and Pearson correlation analysis to solve the problem of feature redundancy. Then, the spatial features are extracted by using a convolutional neural network, and further extracted by fusing Averagepooling and Maxpooling, using attention mechanism to assign different weights to the features, thus reducing the overhead and improving the model performance. At the same time, a Gated Recurrent Unit (GRU) is used to extract the long-distance dependent information features to achieve comprehensive and effective feature learning. Finally, a softmax function is used for classification. The proposed intrusion detection model is evaluated based on the UNSW_NB15, NSL-KDD, and CIC-IDS2017 datasets, and the experimental results show that the classification accuracy reaches 86.25%, 99.69%, 99.65%, which are 1.95%, 0.47% and 0.12% higher than that of the same type of CNN-GRU, and can solve the problems of low classification accuracy and class imbalance well. [ABSTRACT FROM AUTHOR]

Medical diagnosis technology based on convolutional neural networks (CNNs) has achieved good performance. In this study, we collected serum samples from 38 glioma patients and 45 healthy controls and used partial least squares (PLS) analysis to reduce the dimension of the data. Different levels of noise were added to the reduced data onto data augmentation, and the AlexNet, ResNet, and GoogLeNet fine‐tuning models were applied for classification. To evaluate the performance of the models, we used five‐fold cross‐validation. The accuracy rates of the AlexNet, ResNet, and GoogLeNet fine‐tuning models were 98.50%, 98.24%, and 99.50%, respectively. The model with the best classification effect was GoogLeNet. The specificity and sensitivity of this model were 98.98% and 98.48%, respectively. In addition, the area under the receiver operating characteristic (ROC) curve (AUC) of the established diagnostic model was 0.9949. The results showed that the combination of serum Raman spectroscopy and the PLS‐Gaussian‐GoogLeNet model achieved a good diagnostic effect for glioma. This method has high clinical application value and is worthy of further popularization. [ABSTRACT FROM AUTHOR]

Protein fold recognition is critical for studying the structures and functions of proteins. The existing protein fold recognition approaches failed to efficiently calculate the pairwise sequence similarity scores of the proteins in the same fold sharing low sequence similarities. Furthermore, the existing feature vectorization strategies are not able to measure the global relationships among proteins from different protein folds. In this article, we proposed a new computational predictor called DeepSVM-fold for protein fold recognition by introducing a new feature vector based on the pairwise sequence similarity scores calculated from the fold-specific features extracted by deep learning networks. The feature vectors are then fed into a support vector machine to construct the predictor. Experimental results on the benchmark dataset (LE) show that DeepSVM-fold obviously outperforms all the other competing methods. [ABSTRACT FROM AUTHOR]

Monitoring pig drinking has been a topic of interest to pig researchers and producers for many years. However, challenges still remain due to the fact that pigs like to play with drinkers in nursery environments and that the drinking pig is often touching others. These factors negatively influence the performance of camera-based pig drinking detection algorithms. The aim of this study is to investigate a deep learning method based on convolutional neural network (CNN) and long short-term memory (LSTM) to classify drinking and drinker-playing. In the experiment, two pens of pigs were video recorded for 3 days. In video from the first pen, 5400 2 s drinking episodes and 5400 2 s drinker-playing episodes were generated with 80% of these data being allocated as training set and the remaining 20% as validation set. In video from the second pen, 12,000 2 s drinking and drinker-playing episodes were generated as a test set. Firstly, the CNN architecture ResNet50 was used to extract spatial features. These features were input into LSTM framework to extract spatial–temporal features. Through the fully connected layer, the prediction function Softmax was finally used to classify these drinking and drinker-playing episodes. In the test set, the classification accuracy in the body and head regions of interest was 87.2% and 92.5%, respectively. The results indicate that the proposed method can be used to classify pigs' drinking and drinker-playing. These classification results have potential to improve the accuracy of pig drinking detection and help farmers to estimate pig welfare. • A novel video-based deep learning method was used to study pig drinking behaviour. • Long short-term memory was used to extract the spatial–temporal features. • Classification accuracy in the body and head regions of interest was 87.2% and 92.5%. • Drinking behaviour can be detected under conditions of pigs touching and overlapping. • Shortening the region of interest can be used to improve the classification accuracy. [ABSTRACT FROM AUTHOR]

In zero-shot learning, attributes play as a bridge from original images to class labels. Therefore, to achieve accurate zero-shot image classification, we mainly focus on improving attribute prediction accuracy by taking full advantage of prior information about attribute from two aspects. First, we present a new attribute classifier called deep attribute prediction (DeepAP) model by using supervised deep convolutional neural networks (DCNNs), where the attribute label information participates in the training of DCNNs. Unlike common DCNNs that are usually used to extract image features, the constructed DCNNs are used to directly predict attribute values from the original input images. Thus, the designed DeepAP model can serve as the mapping from low-level image features to high-level semantic attributes in the traditional direct attribute prediction (DAP) model. Second, another prior information about attribute, i.e., class-attribute matrix is used to mine the attribute-class correlation with sparse representation coefficients. Since the attribute-class correlation can reflect different contributions of attributes to classification, we use it to define attribute weights and incorporate the idea of weighted attributes into DeepAP to form the deep weighted attribute prediction (DWAP) model. Experiments on three real datasets show that DWAP outperforms the deep attribute network and DAP on attribute prediction and zero-shot image classification. [ABSTRACT FROM AUTHOR]

Automatic sleep staging methods usually extract hand-crafted features or network trained features from signals recorded by polysomnography (PSG), and then estimate the stages by various classifiers. In this study, we propose a classification approach based on a hierarchical neural network to process multi-channel PSG signals for improving the performance of automatic five-class sleep staging. The proposed hierarchical network contains two stages: comprehensive feature learning stage and sequence learning stage. The first stage is used to obtain the feature matrix by fusing the hand-crafted features and network trained features. A multi-flow recurrent neural network (RNN) as the second stage is utilized to fully learn temporal information between sleep epochs and fine-tune the parameters in the first stage. The proposed model was evaluated by 147 full night recordings in a public sleep database, the Montreal Archive of Sleep Studies (MASS). The proposed approach can achieve the overall accuracy of 0.878, and the F1-score is 0.818. The results show that the approach can achieve better performance compared to the state-of-the-art methods. Ablation experiment and model analysis proved the effectiveness of different components of the proposed model. The proposed approach allows automatic sleep stage classification by multi-channel PSG signals with different criteria standards, signal characteristics, and epoch divisions, and it has the potential to exploit sleep information comprehensively. [ABSTRACT FROM AUTHOR]

Highlights • Propose a new super-resolution method based on CSC and CNN. • Use CNN to reconstruct the high resolution image for the low-frequency part. • Propose an ACSC to reconstruct high resolution image for high-frequency part. • Integrate and develope the advantages of the CSC method and CNN. Abstract The convolutional sparse coding-based super-resolution (CSC-SR) method has shown its good performance in single image super-resolution. It divides the low-resolution (LR) image into low-frequency part and the high-frequency part, and reconstructs their corresponding high-resolution (HR) image with bicubic interpolation and convolutional sparse coding (CSC) method, respectively. This paper is devoted to improve the performance of CSC-SR method. As convolutional neural network (CNN) can reveal the mapping relation between the LR image and the HR image for the low-frequency part better, we replace the bicubic interpolation with CNN to reconstruct the HR image for the low-frequency part. In addition, we propose an adaptive CSC method to reconstruct the HR image for the high-frequency part. We name our proposed super-resolution method as hybrid adaptive convolutional sparse coding-based super-resolution (HACSC-SR) method. Many comparison experiments illustrate that our proposed HACSC-SR method is superior to CSC-SR, CNN as well as several existing super-resolution methods. [ABSTRACT FROM AUTHOR]

We study the problem of 3D object detection from RGB-D images so as to achieve localization (i.e., producing a bounding box around the object) and classification (i.e., determining the object category) simultaneously. Its challenges arise from high intra-class variability, illumination change, background clutter and occlusion. To solve this problem, we propose a novel solution that integrates the 2D information (RGB images), the 3D information (RGB-D images) and the object/scene context information together, and call it the Context-Assisted 3D (C3D) method. In the proposed C3D method, we first use a convolutional neural network (CNN) to jointly detect a 3D object in a scene and its scene category. Then, we improve the detection result furthermore with a Conditional Random Field (CRF) model that incorporates the object potential, the scene potential, the scene/object context, the object/object context, and the room geometry. Extensive experiments are conducted to demonstrate that the proposed C3D method achieves the state-of-the-art performance for 3D object detection against the SUN RGB-D benchmark dataset. [ABSTRACT FROM AUTHOR]

An approach that extracts global attributes from outdoor images to facilitate geometric layout labeling is investigated in this work. The proposed Global-attributes Assisted Labeling (GAL) system exploits both local features and global attributes. First, by following a classical method, we use local features to provide initial labels for all super-pixels. Then, we develop a set of techniques to extract global attributes from 2D outdoor images. They include sky lines, ground lines, vanishing lines, etc. Finally, we propose the GAL system that integrates global attributes in the conditional random field (CRF) framework to improve initial labels so as to offer a more robust labeling result. The performance of the proposed GAL system is demonstrated and benchmarked with several state-of-the-art algorithms against a popular outdoor scene layout dataset. [ABSTRACT FROM AUTHOR]

• It is the first time to use computer vision to recognise pig enrichment engagement. • A HSV colour space-based method was proposed to track the object region of interest. • InceptionV3 and LSTM were used to extract discriminative spatial-temporal features. • This algorithm has reference value for classifying similar behaviours of pigs. • Pigs preference to 3 enrichment objects was also determined in this study. As providing objects that pigs prefer can reduce the occurrence of tail-biting and aggression and consequently improve animal welfare, automatic recognition of pigs' engagement with different objects can have practical value. Therefore, aim of this study was to develop a computer vision based approach that utilised a recurrent neural network-based deep learning algorithm to recognise pig enrichment engagement (EE) behaviours and preliminarily determine the preference to objects. Two pig pens were studied. 1 day of video was recorded in pen 1, which generated 2400 1 s EE and 2400 1 s non-EE episodes. 80% of these data was randomly selected as training set and the remaining 20% as validation set. Moreover, 4 days of video were recorded and used as the test set in pen 2. Firstly, the HSV (Hue, Saturation, Value) colour space-based tracking algorithm was developed to locate object region of interest. Secondly, the convolutional neural network (CNN) architecture InceptionV3 was used to extract spatial features from each frame. These features were input into the long short-term memory (LSTM) framework to extract spatial-temporal features from each episode. Through the fully connected layer, the prediction function Softmax was finally used to classify these episodes as EE or non-EE behaviour. In the validation set, the proposed algorithm could recognise EE with blue ball, golden ball and wooden beam with an accuracy of 95.2%, 95.4% and 97.3%, respectively. By shortening the radius of the region of interest into a half of the average length of pig body, the corresponding accuracy could be improved into 96.9%, 97.1% and 97.9%, respectively. In the test set, the proposed algorithm could recognise EE with each of these 3 objects with an accuracy of 96.5%, 96.8% and 97.6%, respectively. The proportion of EE with each of these 3 objects was 75.8%, 6.0% and 18.2%, respectively. These results indicate that the proposed method can be used to recognise EE behaviours of pigs, and halving the radius of the region of interest can improve the recognition accuracy of EE behaviours. Moreover, the preference of pigs to objects based on EE duration were preliminarily determined as blue ball > wooden beam > golden ball. The obtained duration of EE behaviours can help farmers to evaluate the enrichment used and thereby to increase the health and welfare of the pigs in their care. Furthermore, the proposed algorithm has reference value for the classification of the behaviours with similar motion patterns. [ABSTRACT FROM AUTHOR]

• A deep learning method based on CNN and LSTM was used to recognise pig aggression. • This study directly process video episodes rather than individual frames. • CNN features were input into LSTM framework to extract spatial-temporal features. • The proposed method could recognise aggressive episodes with an accuracy of 97.2%. • A frame skipping approach can improve the accuracy to 98.4% and halve running time. Aggression is considered as a major animal welfare problem in commercial pig farming. The aim of this study is to develop a deep learning method based on convolutional neural network (CNN) and long short-term memory (LSTM) to recognise aggressive episodes of pigs. Compared to previous studies of pig behaviours based on deep learning, this study directly process video episodes rather than individual frames. In the experiment, nursery pigs (8/pen) were mixed for 3 days and then 8 h of video was recorded in each day. From these videos, 600 aggressive 2 s-episodes were manually selected and then augmented into 2400 episodes by using horizontal, vertical and diagonal mirroring. From the videos, 2400 non-aggressive 2 s-episodes were also manually selected. 80% of the data were randomly allocated as training set and the remaining 20% as validation set. Firstly, the CNN architecture VGG-16 was used to extract spatial features. These features were then input into LSTM framework to further extract temporal features. Through fully connected layer, the prediction function Softmax was finally used to determine if the current episode is aggression or non-aggression. Using the proposed method, aggressive episodes could be recognised with an accuracy of 97.2%. This result indicates that this method can be used to recognise aggressive episodes of pigs. Additionally, this paper further investigates the validity of this method under the conditions of skipping frames and reducing the episode length. The results show that a frame skipping approach whereby 30 fps is reduced into 15 fps within each 2 s-episode can improve the accuracy into 98.4% and halve the total running time. [ABSTRACT FROM AUTHOR]

Chlorophyll-a (Chl-a) concentration plays a crucial role in monitoring marine phytoplankton, holding significant implications for marine ecosystems and human livelihoods. Remote sensing serves as a valuable method for directly estimating Chl-a concentration and facilitating marine monitoring activities. In order to accurately estimate Chl-a concentration and effectively monitor the marine ecological environment of the Taiwan Strait, this paper introduces a novel neural network model called Convolutional Neural Network with Dual Attention Mechanism Optimization (CNN-CBAM). Additionally, the spectral bands derived from MODIS 500 m imagery are subjected to various processing techniques to generate four distinct sets of feature data. These datasets are subsequently merged with in-situ measurements of Chl-a concentration obtained from buoy stations. Given the limited availability of buoy stations, the temporal continuity of data is leveraged to overcome the spatial distribution constraints. This enables the estimation of Chl-a concentration at various time intervals within the research area of the Taiwan Strait. To evaluate the performance of the model, metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), and R2 are employed. The results demonstrate that the CNN-CBAM model, utilizing MODIS temporal data, effectively retrieves Chl-a concentration in the research area of the Taiwan Strait with high accuracy. For the CNN-CBAM model, the RMSE is 0.34 μg L−1, MAPE is 22%, and R2 is 0.86. In comparison, the traditional empirical model using the Band Ratio (BR) model has an RMSE of 0.61 μg L−1, MAPE of 57%, and R2 of 0.31. The results indicate that the CNN-CBAM model has improved R2 by 0.55 compared to the traditional BR model. Additionally, the RMSE and MAPE have decreased by 0.27 μg L−1 and 35%, respectively. The superiority of the CNN-CBAM model over the traditional BR model is evident, showcasing a substantial enhancement in the accuracy of Chl-a concentration retrieval. Given the challenges posed by the intricate and ever-changing marine environment, particularly the scarcity of in-situ buoy stations, this study offers a practical solution for monitoring Chl-a concentration. It establishes a framework for leveraging satellite-based multispectral sensors to enable large-scale and multi-temporal remote sensing retrieval of Chl-a concentration in the research area of the Taiwan Strait. • The CNN-CBAM model demonstrates exceptional performance in estimating Chl-a concentration. • The CNN-CBAM model excels in mining non-linear data compared to traditional empirical models. • CNN-CBAM outperforms other machine learning models in extracting high-dimensional feature data. • The incorporation of the CBAM module enhances the performance of the CNN model. [ABSTRACT FROM AUTHOR]

The classification of benign and malignant parotid tumors is very crucial for the selection of surgical methods and their prognoses. The wide application of deep learning technology in the field of medical imaging also provides new ideas for the computer-aided diagnosis of parotid gland tumors. In addition, because the pathological types of parotid gland tumors are very complicated and the computed tomography (CT) images of benign and malignant patients are also very similar, some clinicians may misjudge tumors due to a lack of experience, which affects the effect of surgical treatment and prognosis. Therefore, this research proposes using deep learning methods to solve this problem. This study uses the four classic pretraining models of VGG16, InceptionV3, ResNet and DenseNet to classify parotid CT images using transfer learning methods and uses an improved convolutional neural network (CNN) model to classify parotid CT images. The experimental results show that the improved CNN model achieves an accuracy of 97.78%, and its classification performance is better than those of the other four transfer learning methods. It can effectively diagnose benign and malignant parotid tumors and improve the diagnostic accuracy.