Search

The effective prediction of heart disorders is crucial for timely intervention and treatment before a cardiac event occurs. While various machine learning models have been developed for this purpose, many struggle to handle high-dimensional data effectively, limiting their performance. In this work, efforts have been made to enhance the performance and computational efficiency of deep learning classifiers using hyperparameters. The study utilized heart sound data from normal and diseased patients obtained from standard online repositories. The hyperparameter tuned modified CNN-based Inception Network model achieved an accuracy of 99.65% ± 0.23% on the test dataset, along with a sensitivity of 98.8% ± 0.12% and specificity of 98.2% ± 0.15%. Thus the hyperparameter-tuned CNN-based Inception Network model outperformed its counterparts, making it the most effective model for diagnosing heart disorders. [ABSTRACT FROM AUTHOR]

Vehicle exhaust emissions are a significant factor causing urban environmental pollution in China. Anomaly detection in vehicle exhaust remote sensing data presents challenges due to unreasonable interference data that are difficult to manually identify and eliminate. This study addresses the need for advanced methods to accurately and efficiently identify these anomalies. We propose an enhanced convolutional neural network (CNN) Inception algorithm for recognizing vehicle exhaust data, which converts multidimensional exhaust parameters into images for feature extraction and recognition. Utilizing measured data from Xiqing Saida Road in Tianjin, Meishan area in Sichuan, and Gucheng County in Hebei Hengshui, we established a comprehensive dataset focusing on nitrogen oxides, hydrocarbons, and carbon monoxide concentrations. Three CNN algorithms (AlexNet, VGG16, and Inception) were selected for their proven effectiveness in image recognition tasks and their capability to handle complex data patterns. These models were evaluated for their performance in vehicle exhaust recognition. Experimental results demonstrate that all three algorithms can effectively identify abnormal vehicle exhaust data, with the improved Inception CNN achieving the highest accuracy of 98.7%.The findings underscore the superior performance of the Inception network, establishing it as a robust tool for filtering unreasonable data from remote sensing datasets and efficiently recognizing vehicles with substandard exhaust emissions. This research provides a critical foundation for assessing vehicle compliance and contributes significantly to urban air quality monitoring and pollution control initiatives.Highlights: This paper propose an enhanced convolutional neural network (CNN) Inception algorithm for recognizing vehicle exhaust data. This algorithm has strong effectiveness and the ability to handle complex data patterns in image recognition tasks. This algorithm has been successfully applied to recognition of real vehicle exhaust remote sensing data and has obtained high recognition accuracy. [ABSTRACT FROM AUTHOR]

Abstract Remote sensing image object detection is widely used in civil and military fields. The important task is to detect objects such as ships, planes, airports, harbours and so on, and then it can obtain object category and position information. It is of great significance to use remote sensing images to observe the densely arranged and directional targets such as cars and ships parked in parking lots and harbours. The object detection task mainly includes object localization and classification. Remote sensing images contain large number of small objects and dense scenes due to the long shooting distance and wide coverage. Small objects occupy few pixels in the image, and they are easily miss‐detected. In dense scenes, the overlapping part of each object is large, so it is easy to detect objects repeatedly. The traditional small object detection methods deliver low accuracy and take long time. Therefore, object detection is very challenging. We put forward a novel deep learning‐based single shot multibox detector (SSD) model for object detection. First, we propose an improved inception network to optimize SSD to strengthen the small object feature extraction ability (FEA) in the shallow network. Second, the feature pyramid network is modified to enhance the fusion effect. Third, the deep feature fusion module is designed to improve the FEA of the deep network. Finally, the extracted image features are matched with candidate boxes with different aspect ratios to perform object detection and location with different scales. Experiments on DOTA show that the proposed algorithm can adapt to the remote sensing object detection in different backgrounds, and effectively improve the detection effect of remote sensing objects in complex scenes.

Remote sensing image object detection is widely used in civil and military fields. The important task is to detect objects such as ships, planes, airports, harbours and so on, and then it can obtain object category and position information. It is of great significance to use remote sensing images to observe the densely arranged and directional targets such as cars and ships parked in parking lots and harbours. The object detection task mainly includes object localization and classification. Remote sensing images contain large number of small objects and dense scenes due to the long shooting distance and wide coverage. Small objects occupy few pixels in the image, and they are easily miss‐detected. In dense scenes, the overlapping part of each object is large, so it is easy to detect objects repeatedly. The traditional small object detection methods deliver low accuracy and take long time. Therefore, object detection is very challenging. We put forward a novel deep learning‐based single shot multibox detector (SSD) model for object detection. First, we propose an improved inception network to optimize SSD to strengthen the small object feature extraction ability (FEA) in the shallow network. Second, the feature pyramid network is modified to enhance the fusion effect. Third, the deep feature fusion module is designed to improve the FEA of the deep network. Finally, the extracted image features are matched with candidate boxes with different aspect ratios to perform object detection and location with different scales. Experiments on DOTA show that the proposed algorithm can adapt to the remote sensing object detection in different backgrounds, and effectively improve the detection effect of remote sensing objects in complex scenes. [ABSTRACT FROM AUTHOR]

Cardiac-related disorders are rapidly growing throughout the world. Accurate classification of cardiovascular diseases is an important research topic in healthcare. During COVID-19, auscultating heart sounds was challenging as health workers and doctors wear protective clothing, and direct contact with patients can spread the outbreak. Thus, contactless auscultation of heart sound is necessary. In this paper, a low-cost ear contactless stethoscope is designed where auscultation is done with the help of a bluetooth-enabled micro speaker instead of an earpiece. The PCG recordings are further compared with other standard electronic stethoscopes like Littman 3 M. This work is made to improve the performance of deep learning-based classifiers like recurrent neural networks (RNN) and convolutional neural networks (CNN) for different valvular heart problems using tuning of hyperparameters like learning rate of optimizers, dropout rate, and hidden layer. Hyper-parameter tuning is used to optimize the performances of various deep learning models and their learning curves for real-time analysis. The acoustic, time, and frequency domain features are used in this research. The investigation is made on the heart sounds of normal and diseased patients available from the standard data repository to train the software models. The proposed CNN-based inception network model achieved an accuracy of 99.65 ± 0.06% on the test dataset with a sensitivity of 98.8 ± 0.05% and specificity of 98.2 ± 0.19%. The proposed hybrid CNN-RNN architecture attained 91.17 ± 0.03% accuracy on test data after hyperparameter optimization, whereas the LSTM-based RNN model achieved 82.32 ± 0.11% accuracy. Finally, the evaluated results were compared with machine learning algorithms, and the improved CNN-based Inception Net model is the most effective among others. [ABSTRACT FROM AUTHOR]

Deaf and hearing-impaired people always face communication barriers. Non-invasive surface electromyography (sEMG) sensor-based sign language recognition (SLR) technology can help them to better integrate into social life. Since the traditional tandem convolutional neural network (CNN) structure used in most CNN-based studies inadequately captures the features of the input data, we propose a novel inception architecture with a residual module and dilated convolution (IRDC-net) to enlarge the receptive fields and enrich the feature maps, applying it to SLR tasks for the first time. This work first transformed the time domain signal into a time–frequency domain using discrete Fourier transformation. Second, an IRDC-net was constructed to recognize ten Chinese sign language signs. Third, the tandem CNN networks VGG-net and ResNet-18 were compared with our proposed parallel structure network, IRDC-net. Finally, the public dataset Ninapro DB1 was utilized to verify the generalization performance of the IRDC-net. The results showed that after transforming the time domain sEMG signal into the time–frequency domain, the classification accuracy (acc) increased from 84.29% to 91.70% when using the IRDC-net on our sign language dataset. Furthermore, for the time–frequency information of the public dataset Ninapro DB1, the classification accuracy reached 89.82%; this value is higher than that achieved in other recent studies. As such, our findings contribute to research into SLR tasks and to improving deaf and hearing-impaired people's daily lives. [ABSTRACT FROM AUTHOR]

• Adaptively temporal augmentation is combined with momentum contrast into the pre-training. • Micro-expression sequences (new dataset) are interpolated from raw ones in a recursive way. • A shallow model with inflated inception module is designed to alleviate the overfitting problem. • The pre-trained model outperforms other optical flow-based state-of-the-arts through experiments. Micro-expressions (MEs) can reveal the hidden but real emotion and are usually caused spontaneously. However, the characteristics of subtlety and temporariness with the lack of sufficient ME datasets make it hard for recognition. In this paper, we propose an adaptively temporal augmented momentum contrastive learning to alleviate these problems. For the small scale, we pre-train the model on a new interpolated dataset via contrastive learning with momentum contrast (MoCo). For the subtle and rapid facial movements, we augment the temporal dynamics using an adaptive dropout operation to redundant frames. Specifically, we use a recursive way to create a new interpolated dataset from raw datasets firstly. Then we design a shallow model with an inflated inception module as the encoder of the contrastive learning. Afterward, we pre-train the model on the new dataset via momentum contrastive learning. During the pre-training, we propose adaptively temporal augmentation via generative adversarial learning. After the pre-training, we take the encoder out and finetune it for recognition. Finally, we perform extensive experiments and ablation studies on three ME datasets. The results demonstrate the effectiveness of the MoCo-like pre-training and the temporal augmentation for recognition. Moreover, the pre-trained model outperforms other state-of-the-art models based on optical flow. [ABSTRACT FROM AUTHOR]

The nonlinear relation between the spectral information and the corresponding objects (complex physiognomies) makes pixelwise classification challenging for conventional methods. To deal with nonlinearity issues in hyperspectral image classification (HSIC), convolutional neural networks (CNN) are more suitable, indeed. However, fixed kernel sizes make traditional CNN too specific, neither flexible nor conducive to feature learning, thus impacting on the classification accuracy. The convolution of different kernel size networks may overcome this problem by capturing more discriminating and relevant information. In light of this, the proposed solution aims at combining the core idea of 3-D and 2-D inception net with the attention mechanism to boost the HSIC CNN performance in a hybrid scenario. The resulting attention-fused hybrid network (AfNet) is based on three attention-fused parallel hybrid subnets with different kernels in each block repeatedly using high-level features to enhance the final ground-truth maps. In short, AfNet is able to selectively filter out the discriminative features critical for classification. Several tests on HSI datasets provided competitive results for AfNet compared to state-of-the-art models.

Phalaenopsis orchids are one of the most important exporting commodities for Taiwan. Most orchids are planted and grown in greenhouses. Early detection of orchid diseases is crucially valuable to orchid farmers during orchid cultivation. At present, orchid viral diseases are generally identified with manual observation and the judgment of the grower's experience. The most commonly used assays for virus identification are nucleic acid amplification and serology. However, it is neither time nor cost efficient. Therefore, this study aimed to create a system for automatically identifying the common viral diseases in orchids using the orchid image. Our methods include the following steps: the image preprocessing by color space transformation and gamma correction, detection of leaves by a Unet model, removal of non-leaf fragment areas by connected component labeling, feature acquisition of leaf texture, and disease identification by the two-stagemodel with the integration of a randomforestmodel and an inception network (deep learning) model. Thereby, the proposed system achieved the excellent accuracy of 0.9707 and 0.9180 for the image segmentation of orchid leaves and disease identification, respectively. Furthermore, this system outperformed the naked-eye identification for the easily misidentified categories [cymbidium mosaic virus (CymMV) and odontoglossum ringspot virus (ORSV)] with the accuracy of 0.842 using two-stage model and 0.667 by naked-eye identification. This system would benefit the orchid disease recognition for Phalaenopsis cultivation. [ABSTRACT FROM AUTHOR]

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

The methods with multi-sensor data fusion have been a remarkable way to improve the accuracy and robustness of bearing fault diagnosis under complicated conditions. However, most of the existing fusion models or methods belong to single fusion level and simple fusion structure is usually utilized, and the correlation and complementarity of information between multi-sensor data might be easily ignored. In order to improve the performance of fault diagnosis with multi-sensor data fusion, this paper proposes a novel model of multi-layer deep fusion network with attention mechanism (AMMFN). The proposed model consists of a central network and multiple branch networks stacking by Inception networks, and the deep features of each single-sensor data are extracted automatically by the branch networks, and the extracted features of multi-sensor data at different levels are fused with the central network, and then the information interaction between multi-sensor data can be significantly enhanced and the adaptive hierarchical fusion of information can be achieved. Moreover, a fusion strategy based on attention mechanism is designed to extract more correlation information during the fusion of features extracted from multi-sensor data. Extensive experiments are also performed to evaluate the performance of proposed approach, and the comparison results with other methods indicate that the presented method takes higher accuracy and stronger generalization ability. • An intelligent model based on multi-sensor data is developed for bearing fault diagnosis. • A multi-layer fusion network is developed to adaptively fuse the features at different levels. • A fusion strategy based on attention mechanism is introduced to extract more complementary information. • The accuracy of bearing fault diagnosis is significantly improved by fusing multi-sensor signals. [ABSTRACT FROM AUTHOR]

Deaf and hearing-impaired people always face communication barriers. Non-invasive surface electromyography (sEMG) sensor-based sign language recognition (SLR) technology can help them to better integrate into social life. Since the traditional tandem convolutional neural network (CNN) structure used in most CNN-based studies inadequately captures the features of the input data, we propose a novel inception architecture with a residual module and dilated convolution (IRDC-net) to enlarge the receptive fields and enrich the feature maps, applying it to SLR tasks for the first time. This work first transformed the time domain signal into a time–frequency domain using discrete Fourier transformation. Second, an IRDC-net was constructed to recognize ten Chinese sign language signs. Third, the tandem CNN networks VGG-net and ResNet-18 were compared with our proposed parallel structure network, IRDC-net. Finally, the public dataset Ninapro DB1 was utilized to verify the generalization performance of the IRDC-net. The results showed that after transforming the time domain sEMG signal into the time–frequency domain, the classification accuracy (acc) increased from 84.29% to 91.70% when using the IRDC-net on our sign language dataset. Furthermore, for the time–frequency information of the public dataset Ninapro DB1, the classification accuracy reached 89.82%; this value is higher than that achieved in other recent studies. As such, our findings contribute to research into SLR tasks and to improving deaf and hearing-impaired people’s daily lives.

Objective. Most arrhythmias due to cardiovascular diseases alter the heart’s electrical activity, resulting in morphological alterations in electrocardiogram (ECG) recordings. ECG acquisition is a low-cost, non-invasive process and is commonly used for continuous monitoring as a diagnostic tool for cardiac abnormality identification. Our objective is to diagnose twenty-nine cardiac abnormalities and sinus rhythm using varied lead ECG signals. Approach. This work proposes a deep residual inception network with channel attention mechanism (RINCA) for twenty-nine cardiac arrhythmia classification along with normal ECG from multi-label ECG signal with different lead combinations. The RINCA architecture employing the inception-based convolutional neural network backbone uses residual skip connections with the channel attention mechanism. The inception model facilitates efficient computation and prevents overfitting while exploring deeper networks through dimensionality reduction and stacked 1-dimensional convolutions. The residual skip connections alleviate the vanishing gradient problem. The attention modules selectively leverage the temporally significant segments in a sequence and predominant channels for multi-lead ECG signals, contributing to the decision-making. Main results. Exhaustive experimental evaluation on the large-scale ‘PhysioNet/Computing in Cardiology Challenge (2021)’ dataset demonstrates RINCA 's efficacy. On the hidden test data set, RINCA achieves the challenge metric score of 0.55, 0.51, 0.53, 0.51, and 0.53 (ranked 2nd, 5th, 4th, 5th and 4th) for the twelve-lead, six-lead, four-lead, three-lead, and two-lead combination cases, respectively. Significance. The proposed RINCA model is more robust against varied sampling frequency, recording time, and data with heterogeneous demographics than the existing art. The explainability analysis shows RINCA 's potential in clinical interpretations. [ABSTRACT FROM AUTHOR]

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

Knowledge graphs usually consist of billions of triplet facts describing the real world. Although most of the existing knowledge graphs are huge in scale, they are still far from completion. As a result, varieties of knowledge graph embedding approaches have emerged, which have been proven to be an effective and efficient solution for knowledge graph completion. In this paper, we devise a novel knowledge graph embedding model named InterERP, which aims to improve model performance by increasing Inter actions between the embeddings of E ntities, R elations and relation P aths. Specifically, we first introduce the interaction matrix to obtain the interaction embeddings of entities and relations. Then, we employ the Inception network to learn the query embedding, which can further increase the interactions between entities and relations. Furthermore, we resort to logical rules to construct semantic relation paths and are committed to modeling the interactions between different relations in a relation path. The experimental results on four commonly used datasets, demonstrate that the proposed InterERP matches or outperforms the state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

In this work, we propose two Deep Neural Networks, DNN-1 and DNN-2, based on residual Fast-Slow Refined Highway (FSRH) and Global Atomic Spatial Attention (GASA) to effectively recognize and detect actions. The proposed DNN-1 includes a 3D Convolutional Neural Network (3DCNN), Residual FSRH (R_FSRH), reduction layer, and classification layer for action recognition. In action detection of subject-region extraction and classification, the proposed DNN-2 consists of a 3DCNN, region proposal network, R_FSRH, GASA, and classification-localization layer. The 3DCNN takes the front layer to the “Mixed-3c” layer of the pre-trained Inflated 3D (I3D) network as the backbone structure. FSRH is composed of two Refined Highway (RH) units to extract a pair of features from fast and slow actions, where RH has temporal attention from a non-local 3D convolution and an affine transform by a temporal bilinear inception. In R_FSRH, multiple cascaded FSRHs with different residual connections were investigated to determine an effective one. GASA sequentially computes and concatenates the correlation features of an atomic subject and other subjects to effectively discover high-level semantic information. In ablation studies, extensive experiments were conducted to demonstrate the superior performance of the proposed DNN-1 and DNN-2 on five challenging video datasets of JHMDB-21, UCF101-24, Traffic Police (TP), Charades, and AVA. Notably, the proposed DNN-1 shows state-of-the-art performance of 98.6% on UCF101-24 and 98.1% on TP, and DNN-2 exhibits state-of-the-art video-mAP of 27.7% on AVA, and the second best video-mAP of 25.7% on Charades, to the best of our knowledge. Therefore, the DNN-1 and DNN-2 proposed herein can be outstanding context-aware engines for various video understanding applications.

Traumatic Brain Injury (TBI) leads to intracranial hemorrhages (ICH), which is a severe illness resulted in death if it is not properly diagnosed and treated in the earlier stage. Presently, computer tomography (CT) images are widely used by radiologists to identify and locate the regions of ICH. But it is a tedious task and mainly depends on the professional radiologists. This paper develops new deep learning (DL)-based ICH diagnosis and classification (DL-ICH) model using optimal image segmentation with Inception Network. The proposed DL-ICH model involves preprocessing, segmentation, feature extraction, and classification. Firstly, the input data undergo format conversion where the NIfTI files are converted into JPEG format. Next, Kapur's thresholding with an elephant herd optimization (EHO) algorithm called KT-EHO is employed for image segmentation. Then, DL based Inception v4 network is applied as a feature extractor to extract a useful set of features and a multilayer perceptron (MLP) is used for classification. An extensive set of simulations takes place to ensure the effective diagnostic performance of the DL-ICH model and the results are investigated under diverse dimensions. The experimental results achieved better accuracy rate. [ABSTRACT FROM AUTHOR]

Detection of infected kernels is important for Fusarium head blight (FHB) prevention and product quality assurance in wheat. In this study, Raman spectroscopy (RS) and deep learning networks were used for the determination of FHB-infected wheat kernels. First, the RS spectra of healthy, mild, and severe infection kernels were measured and spectral changes and band attribution were analyzed. Then, the Inception network was improved by residual and channel attention modules to develop the recognition models of FHB infection. The Inception–attention network produced the best determination with accuracies in training set, validation set, and prediction set of 97.13%, 91.49%, and 93.62%, among all models. The average feature map of the channel clarified the important information in feature extraction, itself required to clarify the decision-making strategy. Overall, RS and the Inception–attention network provide a noninvasive, rapid, and accurate determination of FHB-infected wheat kernels and are expected to be applied to other pathogens or diseases in various crops.

A new multi-scale deep learning (MDL) framework is proposed and exploited for conducting image interpolation in this paper. The core of the framework is a seeding network that needs to be designed for the targeted task. For image interpolation, a novel attention-aware inception network (AIN) is developed as the seeding network; it has two key stages: 1) feature extraction based on the low-resolution input image; and 2) feature-to-image mapping to enlarge image’s size or resolution. Note that the designed seeding network, AIN, needs to be trained with a matched training dataset at each scale. For that, multi-scale image patches are generated using our proposed pyramid cut, which outperforms the conventional image pyramid method by completely avoiding aliasing issue. After training, the trained AINs are then combined for processing the input image in the testing stage. Extensive experimental simulation results obtained from seven image datasets (comprising 359 images in total) have clearly shown that the proposed MAIN consistently delivers highly accurate interpolated images. [ABSTRACT FROM AUTHOR]

Machine learning and computer vision have driven many of the greatest advances in the modeling of Deep Convolutional Neural Networks (DCNNs). Nowadays, most of the research has been focused on improving recognition accuracy with better DCNN models and learning approaches. The recurrent convolutional approach is not applied very much, other than in a few DCNN architectures. On the other hand, Inception-v4 and Residual networks have promptly become popular among computer the vision community. In this paper, we introduce a new DCNN model called the Inception Recurrent Residual Convolutional Neural Network (IRRCNN), which utilizes the power of the Recurrent Convolutional Neural Network (RCNN), the Inception network, and the Residual network. This approach improves the recognition accuracy of the Inception-residual network with same number of network parameters. In addition, this proposed architecture generalizes the Inception network, the RCNN, and the Residual network with significantly improved training accuracy. We have empirically evaluated the performance of the IRRCNN model on different benchmarks including CIFAR-10, CIFAR-100, TinyImageNet-200, and CU3D-100. The experimental results show higher recognition accuracy against most of the popular DCNN models including the RCNN. We have also investigated the performance of the IRRCNN approach against the Equivalent Inception Network (EIN) and the Equivalent Inception Residual Network (EIRN) counterpart on the CIFAR-100 dataset. We report around 4.53, 4.49 and 3.56% improvement in classification accuracy compared with the RCNN, EIN, and EIRN on the CIFAR-100 dataset respectively. Furthermore, the experiment has been conducted on the TinyImageNet-200 and CU3D-100 datasets where the IRRCNN provides better testing accuracy compared to the Inception Recurrent CNN, the EIN, the EIRN, Inception-v3, and Wide Residual Networks. [ABSTRACT FROM AUTHOR]

Beta‐turn prediction is useful in protein function studies and experimental design. Although recent approaches using machine‐learning techniques such as support vector machine (SVM), neural networks, and K nearest neighbor have achieved good results for beta‐turn prediction, there is still significant room for improvement. As previous predictors utilized features in a sliding window of 4‐20 residues to capture interactions among sequentially neighboring residues, such feature engineering may result in incomplete or biased features and neglect interactions among long‐range residues. Deep neural networks provide a new opportunity to address these issues. Here, we proposed a deep dense inception network (DeepDIN) for beta‐turn prediction, which takes advantage of the state‐of‐the‐art deep neural network design of dense networks and inception networks. A test on a recent BT6376 benchmark data set shows that DeepDIN outperformed the previous best tool BetaTPred3 significantly in both the overall prediction accuracy and the nine‐type beta‐turn classification accuracy. A tool, called MUFold‐BetaTurn, was developed, which is the first beta‐turn prediction tool utilizing deep neural networks. The tool can be downloaded at http://dslsrv8.cs.missouri.edu/~cf797/MUFoldBetaTurn/download.html. [ABSTRACT FROM AUTHOR]

The impact of JPEG compression on deep learning (DL) in image classification is revisited. Given an underlying deep neural network (DNN) pre-trained with pristine ImageNet images, it is demonstrated that, if, for any original image, one can select, among its many JPEG compressed versions including its original version, a suitable version as an input to the underlying DNN, then the classification accuracy of the underlying DNN can be improved significantly while the size in bits of the selected input is, on average, reduced dramatically in comparison with the original image. This is in contrast to the conventional understanding that JPEG compression generally degrades the classification accuracy of DL. Specifically, for each original image, consider its 10 JPEG compressed versions with their quality factor (QF) values from {100,90,80,70,60,50,40,30,20,10}. Under the assumption that the ground truth label of the original image is known at the time of selecting an input, but unknown to the underlying DNN, we present a selector called Highest Rank Selector (HRS). It is shown that HRS is optimal in the sense of achieving the highest Top k accuracy on any set of images for any k among all possible selectors. When the underlying DNN is Inception V3 or ResNet-50 V2, HRS improves, on average, the Top 1 classification accuracy and Top 5 classification accuracy on the whole ImageNet validation dataset by 5.6% and 1.9%, respectively, while reducing the input size in bits dramatically—the compression ratio (CR) between the size of the original images and the size of the selected input images by HRS is 8 for the whole ImageNet validation dataset. When the ground truth label of the original image is unknown at the time of selection, we further propose a new convolutional neural network (CNN) topology which is based on the underlying DNN and takes the original image and its 10 JPEG compressed versions as 11 parallel inputs. It is demonstrated that the proposed new CNN topology, even when partially trained, can consistently improve the Top 1 accuracy of Inception V3 and ResNet-50 V2 by approximately 0.4% and the Top 5 accuracy of Inception V3 and ResNet-50 V2 by 0.32% and 0.2%, respectively. Other selectors without the knowledge of the ground truth label of the original image are also presented. They maintain the Top 1 accuracy, the Top 5 accuracy, or the Top 1 and Top 5 accuracy of the underlying DNN, while achieving CRs of 8.8, 3.3, and 3.1, respectively.

Recent advances in remote sensing have shown great potential for different kind of applications like vegetation and crop supervision. Using hyperspectral images (HSI), this process can be performed over large areas of land, allowing for a fast and non-invasive analysis of variables such as water stress, diseases, type of crops, among many others. In this context, representative spatial-spectral features are crucial to develop the data classification process of hyperspectral remote sensing images. In this way, the use of deep learning networks has shown a remarkable classification performance in many applications as the processing of hyperspectral images. In this paper, a data classification model that use kernel PCA (K-PCA) and inception network architecture to generate deep spatial-spectral features is proposed. The features generated are used to label the different kind of classes into the HSI. The labeling process uses a logistic regression (LR) algorithm. This model was validated using six different datasets. The experiments performed shown that the proposed strategy allows improving the data classification performance regarding the results obtained by the traditional stacked layer model of convolutional neural networks (CNN). Quantitatively speaking, the overall accuracy over all experiments of the proposed model for all tested datasets is greater than 92%. [ABSTRACT FROM AUTHOR]

Electroluminescence (EL) imaging provides high spatial resolution and better identifies micro-defects for inspection of photovoltaic (PV) modules. However, the analysis of EL images could be typically a challenging process due to complex defect patterns and inhomogeneous background structure. In this study, a deep convolutional neural network (CNN) model using residual connections and spatial pyramid pooling (SPP) is proposed for the efficient classification of PV cell defects. The proposed CNN model is built on the Inception-v3 network. In this way, feature maps in inception modules are shared to reuse in deeper layers and the representation ability of features is enriched with the pooling process of the SPP in different sizes. Due to the imbalanced class distribution, offline data augmentation strategies are applied and network performance is further improved. The proposed method is evaluated on a publicly available dataset of 8 classes, of which 7 classes are defective and one class is defect-free images. In the comparative evaluation, while other approaches give accuracy values between 76.49% and 89.17%, this value is increased to 93.59% with the proposed method. The experimental results show that the proposed method exhibits more accurate and robust classification performance compared with other model combinations and CNN models. [ABSTRACT FROM AUTHOR]

About 380 thousand people in the United States die each year because of coronary artery disease. About 2 in 10 deaths from CAD happen in adults less than 65 years old. Early detection of blocked arteries can prevent serious surgeries and decrease the number of deaths. Coronary angiography, as one of the methods for diagnosing the state of the arteries, is a complex and invasive procedure. Neural networks can be used to support physicians in the diagnostic process and improve stenoses detection effectiveness. In this paper, we compare different variants of the Inception Network and the Vision Transformer in a stenosis detection task that, in our case, is a binary classification problem. The data set used in the experiments consists of small fragments extracted from coronary angiography videos. Furthermore, we analyze the impact of different percentages of arteries in fragments without stenosis on model performance and show that the data set configuration plays an important role. As the fragment images are small, we analyzed the Sharpness-Aware Minimization together with the Visual Transformers. We employ explainable AI methods to understand the differences in classification performance between selected models. Our results show that convolutional neural networks generally perform better than transformer-based architectures, but perform slightly worse if the Visual Transformer model is supported by the Sharpness-Aware Minimization method. While increasing the percentage of arteries in non-stenosis class, the F1 median decreases by 2.3, 5.4, and 2.2 for the Inception, ViT, and SAM-ViT models. The F1 mean decreases by 3.4, 4.8, and 2.3. The SAM-ViT is more stable compared to other models and based on the F1 mean outperforms other models. This indicates the usefulness of ViT in medical applications, especially in the analysis of coronary angiography. • Comparison of different models in different negative sample scenarios. • Sharpness-Aware Minimization with Vision Transformer used for stenosis detection. • Insight into the RoI and features importance in detection of stenosis. [ABSTRACT FROM AUTHOR]

• A novel hybrid deep CNN has been proposed using Inception module and SR-block, named ISR-CNN. • All transition state data is discarded so that the data used to construct the dataset can be easily and correctly labelled. • The proposed method classifies chatter into slight and severe chatter based on small probability assumptions. • The cutting conditions in the test set are different from that in the training set while the cutting conditions in the validation set are the same as that in the training set. • The proposed method has higher chatter detection accuracy and generalisation capability. Unstable chatter seriously reduces the quality of machined workpiece and machining efficiency. In order to improve productivity, on-line chatter detection has attracted much interest in the past decades. Nevertheless, traditional methods are inevitably flawed due to the manually extracted features. Deep learning methods possess outstanding feature learning and classification capabilities, but the generalisation and accuracy are severely affected by the labelling and training of data. To address this, this paper proposed a novel hybrid deep convolutional neural network method combining an Inception module and a Squeeze-and-Excitation ResNet block (SR-block), namely ISR-CNN. The Inception module can automatically extract multi-scale features of cutting force signal to enrich the feature map. The SR-block can assign weights to different feature channels, thus suppressing useless feature maps and improving the model accuracy. Meanwhile, the introduction of SR-block also reduces the risk of gradient disappearance and speeds up the training of network. The generalisation and accuracy of the model is guaranteed by combining the two modules without training with transition state data. Milling tests were carried out on a wedge-shaped workpiece using different cutting parameters and tool overhang lengths to verify the accuracy and generalisability of the proposed method. The results showed that the proposed method outperforms other methods by achieving classification accuracy of on the validation and test sets 100% and 97.8%, respectively. In comparison to existing methods, the proposed method can correctly identify each machining state, including the transition states. Furthermore, the proposed method identifies the onset of chatter earlier than other methods, which is beneficial for chatter suppression. [ABSTRACT FROM AUTHOR]

At present times, unmanned aerial vehicles (UAVs) have become popular owing to the advantages such as scalable, versatile, autonomous, and inexpensive. The aerial image classification process has attracted more attention among the UAV-enabled surveillance system facilitating objective recognition, tracking process and considerably improving the visual surveillance results. This paper introduces effective deep learning-based aerial image classification using Inception with Residual Network v2 and multilayer perceptron (DLIRV2-MLP). The presented DLIRV2-MLP method includes a series of operations such as pre-processing, feature extraction, and classification. Primarily, the UAVs are used for the collection of aerial images by the use of imaging sensors exist in it. Followed by, the Inception with ResNet v2 based feature extractor is utilized to derive a useful set of feature vectors from the aerial image. Finally, the classification of aerial images using the derived feature vectors takes place by the use of MLP model. Here, the MLP classifier is used as the final layer of the Inception with ResNet v2 model in place of the softmax layer to improve the classification performance. The efficacy of the DLIRV2-MLP method is validated using a benchmark aerial image dataset and the resultant experimental values highlighted the effectiveness of the DLIRV2-MLP method with 90% and above in accuracy and precision. The problem of existing method faces some issues in computation whereas the aerial image classification with DLIRV2-MLP gives efficient computation while comparing with the existing method. [ABSTRACT FROM AUTHOR]

Phalaenopsis orchids are one of the most important exporting commodities for Taiwan. Most orchids are planted and grown in greenhouses. Early detection of orchid diseases is crucially valuable to orchid farmers during orchid cultivation. At present, orchid viral diseases are generally identified with manual observation and the judgment of the grower's experience. The most commonly used assays for virus identification are nucleic acid amplification and serology. However, it is neither time nor cost efficient. Therefore, this study aimed to create a system for automatically identifying the common viral diseases in orchids using the orchid image. Our methods include the following steps: the image preprocessing by color space transformation and gamma correction, detection of leaves by a U-net model, removal of non-leaf fragment areas by connected component labeling, feature acquisition of leaf texture, and disease identification by the two-stage model with the integration of a random forest model and an inception network (deep learning) model. Thereby, the proposed system achieved the excellent accuracy of 0.9707 and 0.9180 for the image segmentation of orchid leaves and disease identification, respectively. Furthermore, this system outperformed the naked-eye identification for the easily misidentified categories [cymbidium mosaic virus (CymMV) and odontoglossum ringspot virus (ORSV)] with the accuracy of 0.842 using two-stage model and 0.667 by naked-eye identification. This system would benefit the orchid disease recognition for Phalaenopsis cultivation., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Tsai, Huang, Wu, Lin, Lee, Yu, Chan and Jan.)

Detection of infected kernels is important for Fusarium head blight (FHB) prevention and product quality assurance in wheat. In this study, Raman spectroscopy (RS) and deep learning networks were used for the determination of FHB-infected wheat kernels. First, the RS spectra of healthy, mild, and severe infection kernels were measured and spectral changes and band attribution were analyzed. Then, the Inception network was improved by residual and channel attention modules to develop the recognition models of FHB infection. The Inception–attention network produced the best determination with accuracies in training set, validation set, and prediction set of 97.13%, 91.49%, and 93.62%, among all models. The average feature map of the channel clarified the important information in feature extraction, itself required to clarify the decision-making strategy. Overall, RS and the Inception–attention network provide a noninvasive, rapid, and accurate determination of FHB-infected wheat kernels and are expected to be applied to other pathogens or diseases in various crops. [ABSTRACT FROM AUTHOR]

• The reccurence plots (RP) suffer from the multi-scale and tendency confusion problems. • Multi-scale signed recurrence plots (MSRP) are proposed to handle the defects of RP for better representation abilities. • Existing time series classification (TSC) networks cannot adapt to the scale variability of MSRP effectively. • The inception fully convolutional networks (IFCN) are proposed, which better extract multi-scale features from MSRP images. • Our proposed MSRP-IFCN achieves superior performance on 85 UCR datasets. Inspired by the great success of deep neural networks in image classification, recent works use Recurrence Plots (RP) to encode time series as images for classification. RP provide rich texture information and construct long-term time correlations, which are effective supplements to the networks. However, RP cannot handle the scale and length variability of sequences. Moreover, RP have serious tendency confusion problem. They cannot represent the upward and downward trends of sequences effectively. In addition to the defects of RP, existing time series classification (TSC) networks cannot adapt to the various scales of discriminative regions of time series effectively. To tackle these problems, this paper proposes a method, named MSRP-IFCN. It is composed of two submodules, the Multi-scale Signed RP (MSRP) and the Inception Fully Convolutional Network (IFCN). MSRP are proposed to handle the defects of RP. They comprise three components, namely the multi-scale RP, the asymmetric RP and the signed RP. We first use the multi-scale RP to enrich the scales of images. Then, the asymmetric RP are constructed to represent long sequences. Finally, the signed RP images are obtained by multiplying the designed sign masks to remove the tendency confusion. Besides, IFCN is proposed to enhance the existing TSC networks in multi-scale feature extraction. By introducing the modified Inception modules, IFCN obtains extensive receptive fields and better extracts multi-scale features from the MSRP images. Experimental results on 85 UCR datasets indicate the superior performance of MSRP-IFCN. The visualization results further demonstrate the effectiveness of our method. [ABSTRACT FROM AUTHOR]

• High diagnostic accuracy be obtained with smaller data sets. • Without any manual time-consuming feature extraction process. • A strong self-adaptive ability, and suitable for practical application. • Using the dynamic routing algorithm in fault diagnosis of rotating machinery. The bearings fault diagnosis is essential for the maintenance and reliability of rotating machinery. Bearings pitting is one of the most common fault types of rotating machinery. However, due to the complex working conditions of bearings, it is challenging to diagnose the pitting faults in bearings inner and outer rings at different speeds. In this paper, an improved one-dimensional inception capsule network (IICN) is proposed to solve the problem of bearing pitting fault diagnosis under complex working conditions. Firstly, the raw bearing vibration signal is processed using the improved Inception network. The function of the stage is to approximate an optimal local sparse structure with a simple dense substructure for bearing healthy state feature extraction. And then inputs concatenated features to the primary capsule layer and the routing capsule layer. The inputs are mapped to feature vector space and weighted by the dynamic routing algorithm. The dynamic routing algorithm encodes the significant spatial relationship between low-level features and upper-level features. The Euclidean length of each capsule vector is the probability of belonging to this bearing healthy condition. In order to validate the effectiveness of the IICN method, bearings pitting experiments at different speeds were designed. The raw bearings vibration signal data under six different health conditions are collected, and the effectiveness of the IICN method is verified. Experimental results show that the IICN method can effectively distinguish different degrees of bearing pitting fault at different speeds, and its diagnostic accuracy is superior to other advanced deep learning methods. [ABSTRACT FROM AUTHOR]

The impact of JPEG compression on deep learning (DL) in image classification is revisited. Given an underlying deep neural network (DNN) pre-trained with pristine ImageNet images, it is demonstrated that, if, for any original image, one can select, among its many JPEG compressed versions including its original version, a suitable version as an input to the underlying DNN, then the classification accuracy of the underlying DNN can be improved significantly while the size in bits of the selected input is, on average, reduced dramatically in comparison with the original image. This is in contrast to the conventional understanding that JPEG compression generally degrades the classification accuracy of DL. Specifically, for each original image, consider its 10 JPEG compressed versions with their quality factor (QF) values from { 100 , 90 , 80 , 70 , 60 , 50 , 40 , 30 , 20 , 10 } . Under the assumption that the ground truth label of the original image is known at the time of selecting an input, but unknown to the underlying DNN, we present a selector called Highest Rank Selector (HRS). It is shown that HRS is optimal in the sense of achieving the highest Top k accuracy on any set of images for any k among all possible selectors. When the underlying DNN is Inception V3 or ResNet-50 V2, HRS improves, on average, the Top 1 classification accuracy and Top 5 classification accuracy on the whole ImageNet validation dataset by 5.6% and 1.9%, respectively, while reducing the input size in bits dramatically—the compression ratio (CR) between the size of the original images and the size of the selected input images by HRS is 8 for the whole ImageNet validation dataset. When the ground truth label of the original image is unknown at the time of selection, we further propose a new convolutional neural network (CNN) topology which is based on the underlying DNN and takes the original image and its 10 JPEG compressed versions as 11 parallel inputs. It is demonstrated that the proposed new CNN topology, even when partially trained, can consistently improve the Top 1 accuracy of Inception V3 and ResNet-50 V2 by approximately 0.4% and the Top 5 accuracy of Inception V3 and ResNet-50 V2 by 0.32% and 0.2%, respectively. Other selectors without the knowledge of the ground truth label of the original image are also presented. They maintain the Top 1 accuracy, the Top 5 accuracy, or the Top 1 and Top 5 accuracy of the underlying DNN, while achieving CRs of 8.8, 3.3, and 3.1, respectively. [ABSTRACT FROM AUTHOR]

• A novel perimeter security abnormal events identification method based on Recurrent Plot algorithm and neural network analysis. • The Recurrent Plot algorithm extracts the motion characteristics of intrusion signals, which improves the accuracy of event identification. • The Recurrent Plot algorithm encoding the signals into two-dimensional images, combined with the Convolutional Neural Networks to improve the efficiency of identification. For establishing an accurate and reliable distributed optical fiber perimeter security system, this paper proposes a novel abnormity detection solution to security using Recurrent Plot (RP) and deep learning technology. Take advantage of the temporal correlation of intrusion signals, we encode the sensing signals into two-dimensional images through the RP algorithm. The RP algorithm can extract the motion characteristics of the signal from the complex time series, and it is robust to instrument noise. These encoded image signatures can reveal the deeper temporal correlation of the intrusion signals' motion. After that, Inception network can adaptively extract the features of these images to complete the accurate identification of a series of noisy intrusion signals. We conducted experiments on three most frequent natural events and three representative man-made intrusion events, including heavy rain, light rain, wind blowing, treading, slapping, and impacting. The results show that the detection accuracy has reached 99.7%. This method can achieve 0.35 s real-time detection in the online detection of abnormal events while ensuring accuracy, providing a new intrusion pattern identification idea for perimeter security. [ABSTRACT FROM AUTHOR]

• Novel approach comprising of data augmentation and deep learning algorithms for improved and reliable malignancy detection of mediastinal lymph nodes (MLNs) in CT images. • Comparison between various state-of-art GAN techniques like DCGAN, CGAN, ACGAN, Info GAN, and WGAN for application in data augmentation. • Comparison between proposed Inception networks (Inception -v4 and Inception-ResNet-v2) and proven DenseNet network for severity detection of MLNs. • Proposed approach achieved best evaluation parameters were, sensitivity = 93.65%, specificity = 96.67%, accuracy = 94.95%, AUC = 95% for severity detection of MLNs. In lung cancer, the determination of mediastinal lymph node (MLN) status as benign or malignant influence treatment planning and survival rate. Invasive pathological tests for the classification of MLNs into benign and malignant have various shortcomings like painfulness, the risk associated with anesthesia, and depends to a large extent on skillset and preferences of the surgeon performing the test. Hence, computer-aided system for MLNs severity detection has been explored widely by the researchers. Very recently, in our earlier concluded work on non-invasive method for MLNs differential diagnosis in computed tomography (CT) images, combination of different data augmentation approaches and state-of-art fully convolutional network (FCN) were implemented to enhance the performance of malignancy detection. However, the performance of FCN network were highly depended on the selection of appropriate data augmentation approach and control of their hyperparameters. Moreover, a standard practice to get hierarchical features in convolutional neural network (CNN) models requires deeper stacking of layers. This leads to an increase in number of trainable parameters which prone to overfitting of the network. In view of the above mention limitations, in this paper, authors have proposed an approach that includes: 1) Generative Adversarial Network (GAN) for data augmentation, and 2) Inception network for malignancy detection. Unlike conventional data augmentation strategy, GAN based augmentation approach generates data that correlates to original data distribution. In the case of Inception based model, it uses multiple size kernels with factorized convolution for hierarchical feature extraction. This helps to a significant reduction in trainable parameters and the problem of overfitting. In this paper, experiments with different GAN approaches, as well as with different Inception architectures, are conducted to evaluate and justify the selection of appropriate GAN and Inception architecture, respectively for MLNs severity detection. The proposed approach achieves superior results with an average accuracy, sensitivity, specificity, and area under curve of 94.95%, 93.65%, 96.67%, and 95%, respectively. The obtained results validate the usefulness of GANs for data augmentation in the differential diagnosis of benign and malignant MLNs. The proposed Inception network based classifier for malignancy detection shows promising results compared to all investigated methods presented in various literature. [ABSTRACT FROM AUTHOR]