Residual Convolutional Neural Network Model Research Articles

A modified deep learning method for Alzheimer’s disease detection based on the facial submicroscopic features in mice

Alzheimer’s disease (AD) is a chronic disease among people aged 65 and older. As the aging population continues to grow at a rapid pace, AD has emerged as a pressing public health issue globally. Early detection of the disease is important, because increasing evidence has illustrated that early diagnosis holds the key to effective treatment of AD. In this work, we developed and refined a multi-layer cyclic Residual convolutional neural network model, specifically tailored to identify AD-related submicroscopic characteristics in the facial images of mice. Our experiments involved classifying the mice into two distinct groups: a normal control group and an AD group. Compared with the other deep learning models, the proposed model achieved a better detection performance in the dataset of the mouse experiment. The accuracy, sensitivity, specificity and precision for AD identification with our proposed model were as high as 99.78%, 100%, 99.65% and 99.44%, respectively. Moreover, the heat maps of AD correlation in the facial images of the mice were acquired with the class activation mapping algorithm. It was proven that the facial images contained AD-related submicroscopic features. Consequently, through our mouse experiments, we validated the feasibility and accuracy of utilizing a facial image-based deep learning model for AD identification. Therefore, the present study suggests the potential of using facial images for AD detection in humans through deep learning-based methods.

Accurate and robust auto-segmentation of head and neck organ-at-risks based on a novel CNN fine-tuning workflow.

Obvious inconsistencies in auto-segmentations exist among various AI software. In this study, we have developed a novel convolutional neural network (CNN) fine-tuning workflow to achieve precise and robust localized segmentation. The datasets include Hubei Cancer Hospital dataset, Cetuximab Head and Neck Public Dataset, and Québec Public Dataset. Seven organs-at-risks (OARs), including brain stem, left parotid gland, esophagus, left optic nerve, optic chiasm, mandible, and pharyngeal constrictor, were selected. The auto-segmentation results from four commercial AI software were first compared with the manual delineations. Then a new multi-scale lightweight residual CNN model with an attention module (named as HN-Net) was trained and tested on 40 samples and 10 samples from Hubei Cancer Hospital, respectively. To enhance the network's accuracy and generalization ability, the fine-tuning workflow utilized an uncertainty estimation method for automatic selection of candidate samples of worthiness from Cetuximab Head and Neck Public Dataset for further training. The segmentation performances were evaluated on the Hubei Cancer Hospital dataset and/or the entire Québec Public Dataset. A maximum difference of 0.13 and 0.7mm in average Dice value and Hausdorff distance value for the seven OARs were observed by four AI software. The proposed HN-Net achieved an average Dice value of 0.14 higher than that of the AI software, and it also outperformed other popular CNN models (HN-Net: 0.79, U-Net: 0.78, U-Net++: 0.78, U-Net-Multi-scale: 0.77, AI software: 0.65). Additionally, the HN-Net fine-tuning workflow by using the local datasets and external public datasets further improved the automatic segmentation with the average Dice value by 0.02. The delineations of commercial AI software need to be carefully reviewed, and localized further training is necessary for clinical practice. The proposed fine-tuning workflow could be feasibly adopted to implement an accurate and robust auto-segmentation model by using local datasets and external public datasets.

Object detection in optical imaging of the Internet of Things based on deep learning.

This article endeavors to enhance image recognition technology within the context of the Internet of Things (IoT). A dynamic image target detection training model is established through the convolutional neural network (CNN) algorithm within the framework of deep learning (DL). Three distinct model configurations are proposed: a nine-layer convolution model, a seven-layer convolution model, and a residual module convolution model. Subsequently, the simulation model of CNN image target detection based on optical imaging is constructed, and the simulation experiments are conducted in scenarios of simple and salient environments, complex and salient environments, and intricate micro-environment. By determining the optimal training iterations, comparisons are drawn in terms of precision, accuracy, Intersection Over Union (IoU), and frames per second (FPS) among different model configurations. Finally, an attention mechanism is incorporated within the DL framework, leading to the construction of an attention mechanism CNN target detection model that operates at three difficulty levels: simple, intermediate, and challenging. Through comparative analysis against prevalent target detection algorithms, this article delves into the accuracy and detection efficiency of various models for IoT target detection. Key findings include: (1) The seven-layer CNN model exhibits commendable accuracy and confidence in simple and salient environments, although it encounters certain instances of undetected images, indicating scope for improvement. (2) The residual network model, when employing a loss function comprising both mean square error (MSE) and cross entropy, demonstrates superior performance in complex and salient environments, manifesting high precision, IoU, and accuracy metrics, thereby establishing itself as a robust detection model. (3) Within intricate micro-environments, the residual CNN model, utilizing loss functions of MSE and cross entropy, yields substantial results, with precision, IoU, and FPS values amounting to 0.99, 0.83, and 29.9, respectively. (4) The CNN model enriched with an attention mechanism outperforms other models in IoT target image detection, achieving the highest accuracy rates of 24.86%, 17.8%, and 14.77% in the simple, intermediate, and challenging levels, respectively. Although this model entails slightly longer detection times, its overall detection performance is excellent, augmenting the effectiveness of object detection within IoT. This article strives to enhance image target detection accuracy and speed, bolster the recognition capability of IoT systems, and refine dynamic image target detection within IoT settings. The implications encompass reduced manual recognition costs and the provision of a theoretical foundation for optimizing imaging and image target detection technologies in the IoT context.

A fine-tuning deep residual convolutional neural network for emotion recognition based on frequency-channel matrices representation of one-dimensional electroencephalography

Emotion recognition (ER) plays a crucial role in enabling machines to perceive human emotional and psychological states, thus enhancing human-machine interaction. Recently, there has been a growing interest in ER based on electroencephalogram (EEG) signals. However, due to the noisy, nonlinear, and nonstationary properties of electroencephalography signals, developing an automatic and high-accuracy ER system is still a challenging task. In this study, a pretrained deep residual convolutional neural network model, including 17 convolutional layers and one fully connected layer with transfer learning technique in combination frequency-channel matrices (FCM) of two-dimensional data based on Welch power spectral density estimate from the one-dimensional EEG data has been proposed for improving the ER by automatically learning the underlying intrinsic features of multi-channel EEG data. The experiment result shows a mean accuracy of 93.61 ± 0.84%, a mean precision of 94.70 ± 0.60%, a mean sensitivity of 95.13 ± 1.02%, a mean specificity of 91.04 ± 1.02%, and a mean F1-score of 94.91 ± 0.68%, respectively using 5-fold cross-validation on the DEAP dataset. Meanwhile, to better explore and understand how the proposed model works, we noted that the ranking of clustering effect of FCM for the same category by employing the t-distributed stochastic neighbor embedding strategy is: softmax layer activation is the best, the middle convolutional layer activation is the second, and the early max pooling layer activation is the worst. These findings confirm the promising potential of combining deep learning approaches with transfer learning techniques and FCM for effective ER tasks.

Solving the identification problems of Bolete origins based on multiple data processing: Take Boletus bainiugan as an example

Wild porcini mushrooms have high nutritional value and are a kind of medicinal food. It is rich in protein and micronutrients needed by the human body, which is popular among consumers. In this paper, we take the problem of origin traceability of Boletus bainiugan as an example and use multiple data processing methods (preprocessing, feature extraction, and data fusion strategies) to establish partial least squares discriminant analysis and support vector machine models. In addition, a deep learning model (3DCOS-ResNet: three-dimensional correlation spectra combined with residual convolutional neural network model) is built to compare with the above two models. It was found that 3DCOS-ResNet was the best model to solve the Boletus baniugan origin traceability problem. Compared to the chemometrics model, it does not require complex processing of the data. The ideal identification effect can be achieved by directly utilizing the raw data, and it saves a lot of time and cost.

Improving the classification accuracy of fishes and invertebrates using residual convolutional neural networks

Abstract The visibility of fishes and invertebrates is highly impacted by the complexity of the environment. Images acquired in underwater environments suffer from blurriness and low contrast. This results in a low classification accuracy. To address this problem, this study uses a pre-trained Resnet50 neural network as the feature extractor, which avoids over-fitting and accuracy saturation while realizing improved feature extraction capabilities. It also proposes an enhancement of the error-minimized random vector functional link (EEMRVFL) neural network, which is used as the classifier in the convolutional neural network (CNN) model instead of the original softmax classifier. EEMRVFL reduces the maximum residual error in each incremental process. The selected hidden nodes are added to the network, which improves the compactness of its structure. The proposed residual CNNs model exhibits improved classification accuracy for underwater image classification compared to existing methods. This is demonstrated experimentally on available datasets such as URPC, LifeCLEF 2015, and Fish4Knowledge with accuracy rates reaching 99.68%, 97.34%, and 99.77%, respectively.

Shaped patch based nonparametric discriminant analysis for hyperspectral image classification through the CNN model

ABSTRACT Convolutional neural network (CNN) has shown superior classification results for hyperspectral image classification. Due to high dimensionality of the hyperspectral image, usually a feature reduction method is applied to the hyperspectral data before feeding it to the CNN model. Principal component analysis is the most widely used approach to this end. However, it has two main disadvantages. It uses the mean square error measure and ignores the class separability. Moreover, it is a pixel-based method, which does not utilize the spatial information. To deal with these weaknesses, the patch-based nonparametric discriminant analysis method is proposed for hyperspectral image feature reduction in this work. However, due to the complex and heterogeneous nature of the hyperspectral images, the use of fixed square patches is not so appropriate. To deal with this difficulty, a shaping approach is introduced for the generation of adaptive patches with consistent shapes in each local region. The input patches of the CNN model are also weighted according to the spectral similarity and class boundaries information. Moreover, a simple residual CNN model is suggested for the classification of the extracted shaped input patches. The efficiency of the proposed framework is assessed in different cases using various evaluation measures on six real hyperspectral images.

Underwater Acoustic Target Recognition Based on Data Augmentation and Residual CNN

In the field of underwater acoustic recognition, machine learning methods rely on a large number of datasets to achieve high accuracy, while the actual collected signal samples are often very scarce, which has a great impact on the recognition performance. This paper presents a recognition method of an underwater acoustic target by the data augmentation technique and the residual convolutional neural network (CNN) model, which is used to expand training samples to improve recognition performance. As a representative model in residual CNN, the ResNet18 model is used for recognition. The whole process mainly includes mel-frequency cepstral coefficient (MFCC) feature extraction, data augmentation processing, and ResNet18 model recognition. On the base of the traditional data augmentation, this study used the deep convolutional generative adversarial network (DCGAN) model to realize the expansion of underwater acoustic samples and compared the recognition performance of support vector machine (SVM), common CNN, VGG19, and ResNet18. The recognition results of the MFCC, constant Q transform (CQT), and low-frequency analyzer and recorder (LOFAR) spectrum were also analyzed and compared. Experimental results showed that the recognition accuracy of the MFCC feature was better than that of other features at the same method, and using the data augmentation method could obviously improve the recognition performance. Moreover, the recognition performance of ResNet18 using data enhancement technology was better than that of other models, which was due to the combination of the data expansion advantage of data augmentation technology and the deep feature extracting ability of the residual CNN model. In addition, although this method was used for ship recognition in this paper, it is not limited to this. This method is also applicable to other target voice recognition, such as natural sound and underwater voice biometrics.

Classification of Diabetic Retinopathy Based on Multiscale Hybrid Attention Mechanism and Residual Algorithm

The key of classification diagnosis of diabetic retinopathy lies in the recognition of the features of small lesions, and it is difficult to extract the features of too small lesions by general extraction methods. In order to solve the problem that it is difficult to extract small focus, a hybrid attention mechanism combined with residual convolutional neural network model algorithm is proposed to improve the classification accuracy of diabetic retinopathy. Firstly, a multiscale deep learning network model with hybrid attention is designed, and then, the high-level features of images are extracted by using the network model; finally, after balancing different types of samples by sampling algorithm, the spatial attention and channel attention of the extracted features are enhanced; small-step learning strategy, loss function, and initial parameters are used to optimize the performance of the network model. The classifier based on multiscale hybrid attention network is used to judge the five classifications. Experimental results show that the proposed algorithm can learn more features of small targets and can effectively improve the classification performance of diabetic retina. An experimental test was performed on Kaggle’s publicly available dataset of diabetic retinas, and the classification accuracy was 93.8%, compared to some existing classification models; the method proposed in this paper can achieve better classification results for diabetic retinopathy.

Inter-patient arrhythmia classification with improved deep residual convolutional neural network

Background and Objective: Early detection of arrhythmias has become critical due to the increased mortality from cardiovascular disease, and ECG is an effective tool for diagnosing cardiovascular disease and detecting arrhythmias. Classification based on ECG signal segments is more suitable for clinical application.Methods: An improved deep residual convolutional neural network is proposed to classify arrhythmias automatically. Firstly, the overlapping segmentation method is used to segment the ECG signals in the MIT-BIH database into segments of 5 seconds in length to overcome the imbalance between classes, and these segments of the ECG signals are re-labeled. Then the discrete wavelet transform (DWT) is used to denoise these segments and the improved deep residual convolutional neural network is used for arrhythmia classification. In addition, the focal loss function is used to overcome the imbalanced classification difficulty between classes.Results: The proposed method gives 94.54% sensitivity, 93.33% positive predictivity, and 80.80% specificity for normal segments. For the supraventricular ectopic segment, the proposed method gives 35.22% sensitivity, 65.88% positive predictivity, and 98.83% specificity. For the ventricular ectopic segment, the proposed method gives 88.35% sensitivity, 79.86% positive predictivity, and 94.92% specificity.Conclusion: The results of this study indicate that the proposed improved deep residual convolutional neural network model trained by the training set obtained by using the overlapping segmentation method is comparable to a classical method and three state-of-art methods. In addition, the classification performance of the network model trained by focal loss as the loss function is further improved.

Hyperspectral Image Classification Based on Multi-Scale Residual Network with Attention Mechanism

In recent years, image classification on hyperspectral imagery utilizing deep learning algorithms has attained good results. Thus, spurred by that finding and to further improve the deep learning classification accuracy, we propose a multi-scale residual convolutional neural network model fused with an efficient channel attention network (MRA-NET) that is appropriate for hyperspectral image classification. The suggested technique comprises a multi-staged architecture, where initially the spectral information of the hyperspectral image is reduced into a two-dimensional tensor, utilizing a principal component analysis (PCA) scheme. Then, the constructed low-dimensional image is input to our proposed ECA-NET deep network, which exploits the advantages of its core components, i.e., multi-scale residual structure and attention mechanisms. We evaluate the performance of the proposed MRA-NET on three public available hyperspectral datasets and demonstrate that, overall, the classification accuracy of our method is 99.82 %, 99.81%, and 99.37, respectively, which is higher compared to the corresponding accuracy of current networks such as 3D convolutional neural network (CNN), three-dimensional residual convolution structure (RES-3D-CNN), and space–spectrum joint deep network (SSRN).

An improved model training method for residual convolutional neural networks in deep learning

Residual convolutional neural network (R-CNN) has become a promising method for image recognition in deep learning applications. The application accuracy, as a key indicator, has a close relationship with filter weights in trained R-CNN models. In order to make filters work at full capacity, we find out that lower relevancy between filters in the same layer promotes higher accuracy for R-CNN applications. Furthermore, we propose an improved R-CNN model training method to acquire a higher accuracy and a better generalization ability. In this paper, the main focus is to control the update of filter weights during model training. The key mechanism is achieved through computing the relevancy between filters in the same layer. The relevancy is quantified by a correlation coefficient, e.g., Pearson Correlation Coefficient (PCC). The mechanism takes a larger probability to utilize the updated filter weights with a lower correlation coefficient, and vice versa. In order to validate our proposal, we construct an experiment through PCC on residual networks. The experiment demonstrates that the improved model training method is a promising mean with better generalization ability and higher recognition accuracy (0.52%-1.83%) for residual networks.

Deep Residual Learning Using Data Augmentation for Median Filtering Forensics of Digital Images

This paper addresses the median filtering forensics for a lossy compressed image with low resolution, which is essential for the identification of fake images and fake videos. A deep residual model with training data augmentation is employed in the proposed method. To solve the dilemma that the low-resolution image is the lack of enough statistical pixels for extracting reliable features, we propose a filter layer to widen the inputs for the convolutional neural network (CNN). First, we perform the high-pass filtering to an image in the filtered layer and stack the multiple filtered residuals into 16-channel feature maps as inputs of CNN. Then, a deep residual CNN model has proposed to self-learn the median filtering traces that are hidden in the JPEG lossy compressed image. To alleviate the over-fitting issue of the deeper CNN model, we employ a data augmentation scheme in the training to increase the diversity of training data and, thus, obtain a more stable median filtering detector. The experimental results demonstrate that the proposed net with training data augmentation outperforms state of the arts in both baseline test and generalization ability test, achieving at least 2% higher in terms of detection accuracy.

Video Superresolution via Motion Compensation and Deep Residual Learning

Video superresolution (SR) techniques are of essential usages for high-resolution display devices due to the current lack of high-resolution videos. Although many algorithms have been proposed, video SR still remains a very challenging inverse problem under different conditions. In this paper, we propose a new method for video SR named motion compensation and residual net (MCResNet). We use optical flow algorithm for motion estimation and motion compensation as a preprocessing step. Then, we employ a novel deep residual convolutional neural network (CNN) to predict a high-resolution image using multiple motion compensated observations. The new residual CNN model preserves the low-frequency contents and facilitates the restoration of high-frequency details. Our method is able to handle large and complex motions adaptively. Extensive experimental results validate that our proposed method outperforms state-of-the-art single-image-based and multi-frame-based algorithms for video SR quantitatively and qualitatively.