Residual Convolutional Neural Network Research Articles

Supervised learning study on ground classification and state recognition of agricultural robots based on multi-source vibration data fusion

In agricultural environments, recognizing the walking ground and state of tracked mobile robots is a complex and challenging task, influenced by clay conditions and other external environmental disturbances. Therefore, this paper proposes a novel data processing method and efficient classifier. Firstly, the noise signals on the left and right sides of the collected robot are averaged, and a time-wavelet-time domain transformation is performed using the Mallat algorithm to achieve nonlinear enhancement of data features and signal denoising. Secondly, the Gramian Angular Summation Fields (GASF) is introduced to transform sequence data into single-channel images, capturing the periodicity and similarity of time series. Next, the images of three sets of sequences are stacked in the channel dimension in RGB format, thus achieving feature fusion of multi-source data. Finally, a supervised learning classifier named Attention-fused Residual Convolutional Neural Network (ANR-CNN) is proposed. Here, the combination of channel and spatial attention mechanisms captures important features in the feature map in both channel and spatial dimensions. The convolutional residual structure enhances feature transmission, improving the model's classification accuracy. Experimental results demonstrate that the proposed data augmentation method effectively enhances model performance, and the classification accuracy of ANR-CNN reaches 92.35%. This implies accurate recognition of the walking ground and state of tracked mobile robots in agricultural environments.

Diagnosing Brain Tumors from MRI images through a Multi-Fused CNN with Auxiliary Layers

In this study, we proposed a novel Multi-Fused Residual Convolutional Neural Network (MFR-CNN) with Auxiliary Fusing Layers (AuxFL) to diagnose various types of brain tumor MRI images. The MFR-CNN was designed to handle four specific cases, namely glioma, meningioma, pituitary, and healthy brain images, obtained from reliable Kaggle databases. Our proposed model integrated three state-of-the-art models into a single feature extraction pipeline, incorporating partially frozen and truncated layers. This strategic fusion enabled the propagation of robust features and improved diagnostic performance without incurring significant computing costs, unlike most existing state-of-the-art models. Moreover, the MFR-CNN effectively mitigated overfitting and performance saturation issues, providing a notable advantage over models lacking these components. Upon evaluation, our proposed model achieved an outstanding accuracy of 94%, surpassing the efficiency and accuracy of conventionally trained DCNNs. Notably, the MFR-CNN demonstrated potential in enhancing brain tumor diagnosis more cost-efficiently than ensembles and outperforming conventional pre-trained and fine-tuned DCNNs. In conclusion, the proposed MFR-CNN with AuxFL and FuRB exhibits promising capabilities to improve the diagnosis of brain tumors, offering better cost-efficiency and accuracy compared to existing methods.

A 4-channelled hazy image input generation and deep learning-based single image dehazing

The images in foggy weather are often degraded which may affect the object detection process. Hence, in recent times, several schemes have been designed to eliminate the haze effect and enhance the performance of computer vision systems. Although these methods reduce the haze effect, they also produce over-saturation and over-degradation on most occasions. These problems occur due to inadequate utilization of the regional haze properties during the dehazing process. Therefore, a new dehazing model is proposed in this paper to address the issues faced by existing dehazing algorithms. The residual convolutional neural network (RCNN) with 14 layers is developed in the present work. In this model, features to indicate the extent of haze are extracted using the regional characteristics of a hazy image and are used as the fourth channel along with the three colour channels of a hazy image. The proposed RCNN is trained with 4-channel hazy images as input and their corresponding haze-free images as output. The results of our model are obtained with several images taken from three different datasets. Through the experimentation, quantitative and qualitative improvements are observed when compared to state-of-the-art dehazing methods.

Rotating Machinery Fault Diagnosis with Limited Multisensor Fusion Samples by Fused Attention-Guided Wasserstein GAN

As vital equipment in modern industry, the health state of rotating machinery influences the production process and equipment safety. However, rotating machinery generally operates in a normal state most of the time, which results in limited fault data, thus greatly constraining the performance of intelligent fault diagnosis methods. To solve this problem, this paper proposes a novel fault diagnosis method for rotating machinery with limited multisensor fusion samples based on the fused attention-guided Wasserstein generative adversarial network (WGAN). Firstly, the dimensionality of collected multisensor data is reduced to three channels by principal component analysis, and then the one-dimensional data of each channel are converted into a two-dimensional pixel matrix, of which the RGB images are obtained by fusing the three-channel two-dimensional images. Subsequently, the limited RGB samples are augmented to obtain sufficient samples utilizing the fused attention-guided WGAN combined with the gradient penalty (FAWGAN-GP) method. Lastly, the augmented samples are applied to train a residual convolutional neural network for fault diagnosis. The effectiveness of the proposed method is demonstrated by two case studies. When training samples per class are 50, 35, 25, and 15 on the KAT-bearing dataset, the average classification accuracy is 99.9%, 99.65%, 99.6%, and 98.7%, respectively. Meanwhile, the methods of multisensor fusion and the fused attention mechanism have an average improvement of 1.51% and 1.09%, respectively, by ablation experiments on the WT gearbox dataset.

Automatic Prediction of Peak Optical Absorption Wavelengths in Molecules Using Convolutional Neural Networks.

Molecular design depends heavily on optical properties for applications such as solar cells and polymer-based batteries. Accurate prediction of these properties is essential, and multiple predictive methods exist, from ab initio to data-driven techniques. Although theoretical methods, such as time-dependent density functional theory (TD-DFT) calculations, have well-established physical relevance and are among the most popular methods in computational physics and chemistry, they exhibit errors that are inherent in their approximate nature. These high-throughput electronic structure calculations also incur a substantial computational cost. With the emergence of big-data initiatives, cost-effective, data-driven methods have gained traction, although their usability is highly contingent on the degree of data quality and sparsity. In this study, we present a workflow that employs deep residual convolutional neural networks (DR-CNN) and gradient boosting feature selection to predict peak optical absorption wavelengths (λmax) exclusively from SMILES representations of dye molecules and solvents; one would normally measure λmax using UV-vis absorption spectroscopy. We use a multifidelity modeling approach, integrating 34,893 DFT calculations and 26,395 experimentally derived λmax data, to deliver more accurate predictions via a Bayesian-optimized gradient boosting machine. Our approach is benchmarked against the state of the art that is reported in the scientific literature; results demonstrate that learnt representations via a DR-CNN workflow that is integrated with other machine learning methods can accelerate the design of molecules for specific optical characteristics.

A novel fusion framework of deep bottleneck residual convolutional neural network for breast cancer classification from mammogram images.

With over 2.1 million new cases of breast cancer diagnosed annually, the incidence and mortality rate of this disease pose severe global health issues for women. Identifying the disease's influence is the only practical way to lessen it immediately. Numerous research works have developed automated methods using different medical imaging to identify BC. Still, the precision of each strategy differs based on the available resources, the issue's nature, and the dataset being used. We proposed a novel deep bottleneck convolutional neural network with a quantum optimization algorithm for breast cancer classification and diagnosis from mammogram images. Two novel deep architectures named three-residual blocks bottleneck and four-residual blocks bottle have been proposed with parallel and single paths. Bayesian Optimization (BO) has been employed to initialize hyperparameter values and train the architectures on the selected dataset. Deep features are extracted from the global average pool layer of both models. After that, a kernel-based canonical correlation analysis and entropy technique is proposed for the extracted deep features fusion. The fused feature set is further refined using an optimization technique named quantum generalized normal distribution optimization. The selected features are finally classified using several neural network classifiers, such as bi-layered and wide-neural networks. The experimental process was conducted on a publicly available mammogram imaging dataset named INbreast, and a maximum accuracy of 96.5% was obtained. Moreover, for the proposed method, the sensitivity rate is 96.45, the precision rate is 96.5, the F1 score value is 96.64, the MCC value is 92.97%, and the Kappa value is 92.97%, respectively. The proposed architectures are further utilized for the diagnosis process of infected regions. In addition, a detailed comparison has been conducted with a few recent techniques showing the proposed framework's higher accuracy and precision rate.

A convolutional neural-network-based diagnostic framework for industrial bearing

Abstract. The problem of industrial bearing health monitoring and fault diagnosis has recently been a popular research topic. Extracting sufficient features from the input raw vibration signals and mapping them to the most likely fault labels is the essence of bearing fault diagnosis. This study proposes a novel framework for bearing defect diagnostics by merging dilated residual convolutional neural networks and attention mechanisms. In this framework, multiple parallel dilated convolutional networks can automatically learn rich fault features at each scale from vibration signals. Simultaneously, the attention approach boosts fault-related features and suppresses irrelevant ones, improving fault detection performance and generalization. According to the experimental results of two different bearing datasets, the framework achieves a higher accuracy and can accurately identify various types of faults.

A zero-shot fault attribute transfer learning method for compound fault diagnosis of power circuit breakers

Diagnosis of compound mechanical faults for power circuit breakers (CBs) is a challenging task. In traditional fault diagnosis methods, however, all fault types need to be collected in advance for the training of diagnosis model. Such processes have poor generalization capabilities for industrial scenarios with no or few data when faced with new faults. In this study, we propose a novel zero-shot learning method named DSR-AL to address this problem. An unsupervised neural network, namely, depthwise separable residual convolutional neural network, is designed to directly learn features from 3D time-frequency images of CB vibration signals. Then we build fault attribute learners (ALs), for transferring fault knowledge to the target faults. Finally, the ALs are used to predict the attribute vector of the target faults, thus realizing the recognition of previously unseen faults. The orthogonal experiments are designed and conducted on real industrial switchgear to validate the effectiveness of the proposed diagnosis framework. Results show that it is feasible to diagnose target faults without using their samples for training, which greatly saves the costs of collecting fault samples. This will help to accurately identify the various faults that may occur during CB’s life cycle, and facilitate the application of intelligent fault diagnosis system.

Research on denoising and enhancing methods of medical images based on convolutional neural networks

SummaryIn the process of modern medical diagnosis, medical image‐assisted diagnosis plays a very important role. However, the process of medical image acquisition, will be affected by various types and degrees of noise, and there will be a certain probability of producing strip artifacts, which will interfere with the doctor's diagnosis, analysis, and treatment of diseases to a certain extent. However, the traditional medical image denoising method will cause problems such as image edge blurring and detail loss, and it is difficult to achieve the balance between noise removal and detail information retention. Therefore, denoising medical images and improving the accuracy of denoising as much as possible have very important scientific research significance and clinical application value. Based on this, this article proposes a medical image denoising method based on a double residual convolutional neural network and compares it with traditional medical images denoising methods such as K‐SVD, BM3D, and PNLM3. Experimental results show that the medical image denoising method based on the double residual convolutional neural network proposed in this article has excellent performance.

Light Weight Residual Convolutional Neural Network for Atrial Fibrillation Detection in Single-lead ECG Recordings

Light Weight Residual Convolutional Neural Network for Atrial Fibrillation Detection in Single-lead ECG Recordings

Construction of a thinking model for Literary Writing based on Deep Spatio-Temporal Residual Convolutional Neural Networks

Construction of a thinking model for Literary Writing based on Deep Spatio-Temporal Residual Convolutional Neural Networks

IDRCNN

Early identification of pancreatic ductal adenocarcinoma (PDAC) improves prognosis. Still, it is difficult since lesions are generally smaller and difficult to define on contrast-enhanced computed tomography images (CE-CT). Ineffective PDAC diagnosis has recently been achieved using deep learning models, but the output localized and identified images are of poor quality. This research focuses on small lesions and presents a new, efficient automatic deep-learning network model for PDAC detection. The Improved Deep Residual Convolutional Neural Network (IDRCNN) detects PDAC. The hyperparameter is optimized using the Tunicate Swarm Optimization Algorithm (TSOA) algorithm. A better diagnosis is made due to segmenting the surrounding anatomy structure effects, such as PD. We train a proposed IDRCNN model for segmenting and detecting lesions automatically using CE-CT images. Two more IDRCNN models are trained with the aim of investigating the effects of anatomy integration: (i) segmentation of tumor and pancreas (IDRCNN_TP), and (ii) segmentation of pancreatic Duct (IDRCNN_PD). The three networks' performance was assessed using an external, publicly available test set. Due to its effective classification results, the proposed method produces improved identification results for automated preliminary diagnosis of PDAC in cervical cancer clinics and hospitals. The performance of the proposed method is evaluated using a publicly assessable CT image dataset. It outperforms the existing state-of-the-art methods and achieved 98.67% accuracy, 97.26% recall, 98.52% precision, 97.65% sensitivity, and 98.45% specificity for pancreatic tumor detection.

Dual Channel Residual Learning for Denoising Path Tracing

In this paper, we present a denoising method for path tracing using residual learning with convolutional neural networks (CNNs). Noisy artifacts in path tracing are inherited from insufficient sampling, which often generates over- or under-exposed values when integrating the limited bright or dark samples in a pixel. In this paper, we introduce a dual channel residual learning CNNs which separates the over and under-exposed signals in order to provide an efficient denoising filter for the path tracing rendering. Furthermore, we present an advanced CNN comprised of variable-sized kernels in each convolutional layer. Our CNN detects features in different scales providing an adaptive denoising filter capability which is optimal for extracting various contextual details in a complex scene. The experiments demonstrate that our method generates better visual quality than other compared approaches across various rendering effects.

Design and experimental demonstration of underwater wireless optical communication system based on semantic communication paradigm.

Underwater wireless optical communication (UWOC) has been widely studied as a key technology for ocean exploration and exploitation. However, current UWOC systems neglect semantic information of transmitted symbols, leading to unnecessary consumption of communication resources for transmitting non-essential data. In this paper, we propose and demonstrate a deep-learning-based underwater wireless optical semantic communication (UWOSC) system for image transmission. By utilizing a deep residual convolutional neural network, the semantic information can be extracted and mapped into the transmitted symbols. Moreover, we design a channel model based on long short-term memory network and employ a two-phase training strategy to ensure that the system matches the underwater channel. To evaluate the performance of the proposed UWOSC system, we conduct a series of experiments on an emulated UWOC experimental platform, in which the effects of different turbidity channel environments and bandwidth compression ratios are investigated. Experimental results show that the UWOSC system exhibits superior performance compared to the conventional communication schemes, particularly in challenging channel environments and low bandwidth compression ratios.

An Interpretable Siamese Attention Res‐CNN for Fingerprint Spoofing Detection

In recent years, fingerprint authentication has gained widespread adoption in diverse identification systems, including smartphones, wearable devices, and attendance machines, etc. Nonetheless, these systems are vulnerable to spoofing attacks from suspicious fingerprints, posing significant risks to privacy. Consequently, a fingerprint presentation attack detection (PAD) strategy is proposed to ensure the security of these systems. Most of the previous work concentrated on how to build a deep learning framework to improve the PAD performance by augmenting fingerprint samples, and little attention has been paid to the fundamental difference between live and fake fingerprints to optimize feature extractors. This paper proposes a new fingerprint liveness detection method based on Siamese attention residual convolutional neural network (Res‐CNN) that offers an interpretative perspective to this challenge. To leverage the variance in ridge continuity features (RCFs) between live and fake fingerprints, a Gabor filter is utilized to enhance the texture details of the fingerprint ridges, followed by the construction of an attention Res‐CNN model to extract RCF between the live and fake fingerprints. The model mitigates the performance deterioration caused by gradient disappearance. Furthermore, to highlight the difference in RCF, a Siamese attention residual network is devised, and the ridge continuity amplification loss function is designed to optimize the training process. Ultimately, the RCF parameters are transferred to the model, and transfer learning is utilized to aid its acquisition, thereby assuring the model’s interpretability. The experimental outcomes conducted on three publicly accessible fingerprint datasets demonstrate the superiority of the proposed method, exhibiting remarkable performance in both true detection rate and average classification error rate. Moreover, our method exhibits remarkable capabilities in PAD tasks, including cross‐material experiments and cross‐sensor experiments. Additionally, we leverage Gradient‐weighted Class Activation Mapping to generate a heatmap that visualizes the interpretability of our model, offering a compelling visual validation.

MLCNNwav: Multilevel Convolutional Neural Network With Wavelet Transformations for Sensor-Based Human Activity Recognition

Human activity recognition (HAR) is a rapidly growing field of research that aims to automatically identify and classify human motions and activities from different tracking devices, such as cameras and sensors. One of the most widely used sensor modalities for HAR is the smartphone, which has various sensors, such as gyroscopes, accelerometers, and GPS, that can provide rich information about a person’s movements and actions. HAR applications are essential for the Internet of Things (IoT) and smart home industries. We used the recent advances in deep learning techniques to develop a new HAR model for wearable sensors. The proposed model, MLCNNwav, relies on residual convolutional neural networks and one-dimensional trainable discrete wavelet transform. The multilevel CNN is designed to capture global features, whereas the wavelet transformation enhances the representation and generalization by learning activity-related features. Several deep learning approaches are compared to assess the superiority of the developed model. Four public benchmarks HAR datasets were used for the evaluation. The outcomes confirmed that the developed MLCNNwav recorded high accuracy rates on all datasets.

Spatio-Temporal Convolutional Neural Network for Enhanced Inter Prediction in Video Coding.

This paper presents a convolutional neural network (CNN)-based enhancement to inter prediction in Versatile Video Coding (VVC). Our approach aims at improving the prediction signal of inter blocks with a residual CNN that incorporates spatial and temporal reference samples. It is motivated by the theoretical consideration that neural network-based methods have a higher degree of signal adaptivity than conventional signal processing methods and that spatially neighboring reference samples have the potential to improve the prediction signal by adapting it to the reconstructed signal in its immediate vicinity. We show that adding a polyphase decomposition stage to the CNN results in a significantly better trade-off between computational complexity and coding performance. Incorporating spatial reference samples in the inter prediction process is challenging: The fact that the input of the CNN for one block may depend on the output of the CNN for preceding blocks prohibits parallel processing. We solve this by introducing a novel signal plane that contains specifically constrained reference samples, enabling parallel decoding while maintaining a high compression efficiency. Overall, experimental results show average bit rate savings of 4.07% and 3.47% for the random access (RA) and low-delay B (LB) configurations of the JVET common test conditions, respectively.

Multi-Resolution Convolutional Residual Neural Networks for Monaural Speech Dereverberation

Multi-Resolution Convolutional Residual Neural Networks for Monaural Speech Dereverberation

Parametric seasonal-trend autoregressive neural network for long-term crop price forecasting.

Crop price forecasting is difficult in that supply is not as elastic as demand, therefore, supply and demand should be stabilized through long-term forecasting and pre-response to the price. In this study, we propose a Parametric Seasonal-Trend Autoregressive Neural Network (PaSTANet), which is a hybrid model that includes both a multi-kernel residual convolution neural network model and a Gaussian seasonality-trend model. To compare the performance of the PaSTANet, we used daily data from the Garak market for four crops: onion, radish, Chinese cabbage, and green onion, and performed long-term price forecasts for one year in 2023. The PaSTANet shows good performance on all four crops compared to other conventional statistical and deep learning-based models. In particular, for onion, the (mean absolute error (MAE) for the long-term forecast of 2023 is 107, outperforming the second-best Prophet (152) by 29.6%. Chinese cabbage, radish, and green onion all outperform the existing models with MAE of 2008, 3703, and 557, respectively. Moreover, using the confidence interval, the predicted price was categorized into three intervals: probability, caution, and warning. Comparing the percentage of classified intervals about the true prices in our test set, we found that they accurately detect the large price volatility.

A New Spatial Downscaling Method for Long-Term AVHRR NDVI by Multiscale Residual Convolutional Neural Network

A New Spatial Downscaling Method for Long-Term AVHRR NDVI by Multiscale Residual Convolutional Neural Network