Traditional Convolutional Neural Network Research Articles

Neural Circuit Policies for Virtual Character Control

The development of high-stakes decision-making neural agents that interact with complex environments, such as video games, is an important aspect of AI research with numerous potential applications. Reinforcement learning combined with deep learning architectures (DRL) has shown remarkable success in various genres of games. The performance of DRL is heavily dependent upon the neural networks resides within them. Although these algorithms perform well in offline testing but the performance deteriorates in noisy and sub-optimal conditions, creating safety and security issues. To address these, we propose a hybrid deep learning architecture that combines a traditional convolutional neural network with worm brain-inspired neural circuit policies. This allows the agent to learn key coherent features from the environment and interpret its dynamics. The obtained DRL agent was not only able to achieve an optimal policy quickly, but it was also the most noise-resilient with the highest success rate. Our research indicates that only 20 control neurons (12 inter-neurons and 8 command neurons) are sufficient to achieve competitive results. We implemented and analyzed the agent in the popular video game Doom, demonstrating its effectiveness in practical applications.

Critical Information Mining Network: Identifying Crop Diseases in Noisy Environments

When agricultural experts explore the use of artificial intelligence technology to identify and detect crop diseases, they mainly focus on the research of a stable environment, but ignore the problem of noise in the process of image acquisition in real situations. To solve this problem, we propose an innovative solution called the Critical Information Mining Network (CIMNet). Compared with traditional models, CIMNet has higher recognition accuracy and wider application scenarios. The network has a good effect on crop disease recognition under noisy environments, and can effectively deal with the interference of noise to the recognition effect in actual farmland scenes. Consider that the shape of the leaves can be symmetrical or asymmetrical.First, we introduce the Non-Local Attention Module (Non-Local), which uses a unique self-attention mechanism to fully capture the context information of the image. The module overcomes the limitation of traditional convolutional neural networks that only rely on local features and ignore global features. Global features are particularly important when the image is disturbed by noise. Non-Local improves a more comprehensive visual understanding of crop disease recognition. Secondly, we have innovatively designed a Multi-scale Critical Information Fusion Module (MSCM). The module uses the Key Information Extraction Module (KIB) to dig into the shallow key features in the network deeply. The shallow key features strengthen the feature perception of the model to the noise image through texture and contour information, and then the shallow key features and deep features are fused to enrich the original deep feature information of the network. Finally, we conducted experiments on two public datasets, and the results showed that the accuracy of our model in crop disease identification under a noisy environment was significantly improved. At the same time, our model also showed excellent performance under stable conditions. The results of this study provide favorable support for the improvement of crop production efficiency.

A bearing fault diagnosis approach based on an improved neural network combined with transfer learning

Abstract In modern industrial systems, bearing failures account for 30–40% of industrial machinery faults. Traditional convolutional neural network suffers from gradient vanishing and overfitting, resulting in a poor diagnostic accuracy. To address the issues, a new bearing fault diagnosis approach was proposed based on an improved AlexNet neural network combined with transfer learning. After decomposition and noise-reduction, reconstructed vibration signals were transformed into 2D images, then input into the improved AlexNet for training and follow-up transfer learning. Program auto-tuning and image-enhancing techniques were employed to increase the diagnostic accuracy in this study. The approach was verified with the datasets from Case Western Reserve University (CWRU), Jiangnan University (JNU), and the Association for Mechanical Failure Prevention Technology (MFPT). The results showed that the diagnostic accuracies by normal learning were more than 97% for CWRU and JNU datasets, and 100% for MFPT dataset. After transfer learning, the accuracies all reached above 99.5%. The proposed approach was demonstrated to be able to effectively diagnose the bearing faults.

Bird Species Detection Using Deep Learning

Bird species classification plays a crucial role in various ecological and conservation endeavors. Leveraging deep learning techniques has shown promise in automating this task, offering the potential for efficient and accurate species identification. This paper presents a comprehensive study on the application of the YOLOv5 architecture, a state-of-the-art object detection framework, for bird species classification from images. We explore the effectiveness of YOLOv5 in comparison to traditional convolutional neural networks (CNNs) for bird species identification. Our study involves extensive experimentation with different configurations of YOLOv5, including variations in model size, training strategies, and dataset preprocessing techniques. We evaluate the performance of the YOLOv5-based classification framework on diverse datasets, ranging from publicly available bird image datasets to custom datasets collected through field observations. Additionally, we compare the classification accuracy of YOLOv5 with other popular CNN architectures, such as ResNet and EfficientNet, to assess its efficacy in bird species recognition tasks. Furthermore, we investigate the transferability of pre-trained YOLOv5 models to different bird species datasets and examine the robustness of the models to variations in image quality, background clutter, and occlusions.Our experimental results demonstrate that YOLOv5 offers competitive performance in bird species classification tasks, achieving high accuracy and efficiency. We discuss the strengths and limitations of using YOLOv5 for this application and provide insights into potential avenues for future research.

A novel intelligent fault diagnosis method for commercial vehicle pneumatic braking system

The application of convolutional neural network (CNN) has greatly broaden the application range of intelligent fault diagnosis, which is beneficial to the stability of industrial production. However, traditional CNN based fault diagnosis methods cannot capture the global features. To address the problem, this paper proposed a method called PSO-Conformer to extract features integrating global and local information. First, convolutional layers are uesd for local feature extraction, and then using Encode module of Transformer to extract global features. The diagnostic performance of the model can be significantly improved and the model hyperparameters are optimized using the PSO algorithm. The proposed method is applied to the fault diagnosis of pneumatic brake systems of commercial vehicles and compared with several methods based on deep learning. And this paper use the t-Distributed Stochastic Neighbor Embdedding (TSNE) to visualize the output features of models in two dimensional plane and four evaluation indexes based on confusion matrix are used to evaluate the model. The results show that the diagnostic performance of the proposed method is better than the existing method.

Intelligent Detection of Surface Defects in High-Speed Railway Ballastless Track Based on Self-Attention and Transfer Learning

The detection of ballastless track surface (BTS) defects is a prerequisite for ensuring the safe operation of high-speed railways. Traditional convolutional neural networks fail to fully exploit contextual information and lack global pixel representations. The extensive stacking of convolutions leads deep learning models to play a black-box detection role, lacking interpretability. Due to the current lack of sufficient high-quality surface data for ballastless tracks, it is a severe constraint on the accurate identification of the substructure state in high-speed railways. This paper proposes an intelligent detection method for BTS defects named TrackNet based on self-attention and transfer learning. The method enhances the fusion ability of global features of BTS defects using multihead self-attention. The model’s dependence on extensive defect data is reduced by transferring knowledge from large-scale publicly available datasets. Experimental results demonstrate that compared to advanced Swin Transformer model results, the TrackNet model achieves improvements in average accuracy and F1-score by 5.15% and 5.16%, respectively, on limited test data. The TrackNet model visualizes the decision regions of the model in identifying BTS defects, revealing the black-box recognition mechanism of deep learning models. This research performs engineering applications and provides valuable insights for the multiclass recognition of BTS defects in high-speed railways.

DATFNets-dynamic adaptive assigned transformer network for fire detection

AbstractFires cause severe damage to the ecological environment and threaten human life and property. Although the traditional convolutional neural network method effectively detects large-area fires, it cannot capture small fires in complex areas through a limited receptive field. At the same time, fires can change at any time due to the influence of wind direction, which challenges fire prevention and control personnel. To solve these problems, a novel dynamic adaptive distribution transformer detection framework is proposed to help firefighters and researchers develop optimal fire management strategies. On the one hand, this framework embeds a context aggregation layer with a masking strategy in the feature extractor to improve the representation of low-level and salient features. The masking strategy can reduce irrelevant information and improve network generalization. On the other hand, designed a dynamic adaptive direction conversion function and sample allocation strategy to fully use adaptive point representation while achieving accurate positioning and classification of fires and screening out representative fire samples in complex backgrounds. In addition, to prevent the network from being limited to the local optimum and discrete points in the sample from causing severe interference to the overall performance, designed a weighted loss function with spatial constraints to optimize the network and penalize the discrete points in the sample. The mAP in the three baseline data sets of FireDets, WildFurgFires, and FireAndSmokes are 0.871, 0.909, and 0.955, respectively. The experimental results are significantly better than other detection methods, which proves that the proposed method has good robustness and detection performance.

Stellar Classification with Vision Transformer and SDSS Photometric Images

With the development of large-scale sky surveys, an increasing number of stellar photometric images have been obtained. However, most stars lack spectroscopic data, which hinders stellar classification. Vision Transformer (ViT) has shown superior performance in image classification tasks compared to most convolutional neural networks (CNNs). In this study, we propose an stellar classification network based on the Transformer architecture, named stellar-ViT, aiming to efficiently and accurately classify the spectral class for stars when provided with photometric images. By utilizing RGB images synthesized from photometric data provided by the Sloan Digital Sky Survey (SDSS), our model can distinguish the seven main stellar categories: O, B, A, F, G, K, and M. Particularly, our stellar-ViT-gri model, which reaches an accuracy of 0.839, outperforms traditional CNNs and the current state-of-the-art stellar classification network SCNet when processing RGB images synthesized from the gri bands. Furthermore, with the introduction of urz band data, the overall accuracy of the stellar-ViT model reaches 0.863, further demonstrating the importance of additional band information in improving classification performance. Our approach showcases the effectiveness and feasibility of using photometric images and Transformers for stellar classification through simple data augmentation strategies and robustness analysis of training dataset sizes. The stellar-ViT model maintains good performance even in small sample scenarios, and the inclusion of urz band data reduces the likelihood of misclassifying samples as lower-temperature subtypes.

Multi-scene application of intelligent inspection robot based on computer vision in power plant

As industries develop, the automation and intelligence level of power plants is constantly improving, and the application of patrol robots is also increasingly widespread. This research combines computer vision technology and particle swarm optimization algorithm to build an obstacle recognition model and obstacle avoidance model of an intelligent patrol robot in a power plant respectively. Firstly, the traditional convolutional recurrent neural network is optimized, and the obstacle recognition model of an intelligent patrol robot is built by combining the connection timing classification algorithm. Then, the artificial potential field method optimizes the traditional particle swarm optimization algorithm, and an obstacle avoidance model of an intelligent patrol robot is built. The performance of the two models was tested, and it was found that the highest precision, recall, and F1 values of the identification model were 0.978, 0.974, and 0.975. The highest precision, recall, and F1 values of the obstacle avoidance model were 0.97, 0.96, and 0.96 respectively. The two optimization models designed in this research have better performance. In conclusion, the two models in this study are superior to the traditional methods in recognition effect and obstacle avoidance efficiency, providing an effective technical scheme for intelligent patrol inspection of power plants.

Lightweight Low-Rank Adaptation Vision Transformer Framework for Cervical Cancer Detection and Cervix Type Classification.

Cervical cancer is a major health concern worldwide, highlighting the urgent need for better early detection methods to improve outcomes for patients. In this study, we present a novel digital pathology classification approach that combines Low-Rank Adaptation (LoRA) with the Vision Transformer (ViT) model. This method is aimed at making cervix type classification more efficient through a deep learning classifier that does not require as much data. The key innovation is the use of LoRA, which allows for the effective training of the model with smaller datasets, making the most of the ability of ViT to represent visual information. This approach performs better than traditional Convolutional Neural Network (CNN) models, including Residual Networks (ResNets), especially when it comes to performance and the ability to generalize in situations where data are limited. Through thorough experiments and analysis on various dataset sizes, we found that our more streamlined classifier is highly accurate in spotting various cervical anomalies across several cases. This work advances the development of sophisticated computer-aided diagnostic systems, facilitating more rapid and accurate detection of cervical cancer, thereby significantly enhancing patient care outcomes.

G2ViT: Graph Neural Network-Guided Vision Transformer Enhanced Network for retinal vessel and coronary angiograph segmentation

Blood vessel segmentation is a crucial stage in extracting morphological characteristics of vessels for the clinical diagnosis of fundus and coronary artery disease. However, traditional convolutional neural networks (CNNs) are confined to learning local vessel features, making it challenging to capture the graph structural information and fail to perceive the global context of vessels. Therefore, we propose a novel graph neural network-guided vision transformer enhanced network (G2ViT) for vessel segmentation. G2ViT skillfully orchestrates the Convolutional Neural Network, Graph Neural Network, and Vision Transformer to enhance comprehension of the entire graphical structure of blood vessels. To achieve deeper insights into the global graph structure and higher-level global context cognizance, we investigate a graph neural network-guided vision transformer module. This module constructs graph-structured representation in an unprecedented manner using the high-level features extracted by CNNs for graph reasoning. To increase the receptive field while ensuring minimal loss of edge information, G2ViT introduces a multi-scale edge feature attention module (MEFA), leveraging dilated convolutions with different dilation rates and the Sobel edge detection algorithm to obtain multi-scale edge information of vessels. To avoid critical information loss during upsampling and downsampling, we design a multi-level feature fusion module (MLF2) to fuse complementary information between coarse and fine features. Experiments on retinal vessel datasets (DRIVE, STARE, CHASE_DB1, and HRF) and coronary angiography datasets (DCA1 and CHUAC) indicate that the G2ViT excels in robustness, generality, and applicability. Furthermore, it has acceptable inference time and computational complexity and presents a new solution for blood vessel segmentation.

Intelligent speech elderly rehabilitation learning assistance system based on deep learning and sensor networks

Recently, deep learning has been proved to significantly improve the quality of speech recognition. Convolutional neural networks are often used in speech recognition tasks because of their special network structure and powerful learning function. In order to solve the problem that the traditional convolutional neural network can not reflect the one-dimensional basic attributes of speech signal, this paper proposes to set the number of frames for one dimension of convolution kernel, and use one-dimensional model and two-dimensional convolutional network model for speech recognition. By moving the convolution kernel of time axis and frequency band, it can adapt to the time change of speech signal and maintain the correlation between frequency bands to a great extent. At the same time, this paper also discusses the speech signal preprocessing, feature parameter extraction and regularization algorithm. Due to the lack of hospital resources, the lag of information technology, the poor ability of accompanying and other reasons, the current accompanying service for elderly rehabilitation can not meet the needs of elderly patients. The rapid development and wide use of information technology provide opportunities for the optimization of health services for the elderly. In view of the difficulty of word memory and interpretation in current rehabilitation courses, a VR intelligent teaching system based on intelligent voice technology is developed. The application results show that the system can effectively improve the ability of language expression and word writing. At present, the system of intelligent speech function has not been completed, and lacks speech synthesis function. The next research will focus on the use of speech synthesis technology, in order to realize the man-machine dialogue between people and the system, and show a more real training situation.

RVS-FDSC: A retinal vessel segmentation method with four-directional strip convolution to enhance feature extraction

Retinal vessel segmentation is important to assist ophthalmologists in the diagnosis and treatment of ophthalmic diseases. However, traditional convolutional neural networks (CNNs) are constrained by using square convolution kernels for retinal vessel feature extraction, thereby overlooking the structural characteristics of retinal vessels, which to some extent limits model performance. Consequently, we propose a retinal vessel segmentation method with four-directional strip convolution to enhance feature extraction (RVS-FDSC). Firstly, due to the shape structure of retinal vessels which have branching paths and appear locally as strips, we have constructed a residual strip convolution (RSC) module by using strip convolutions in four different directions: horizontal, vertical, antidiagonal, and main diagonal. These four directions can roughly simulate the branching directions of vessels, making it possible to better extract vessel features. Furthermore, to more comprehensively understand and utilize both local details and global context, RVS-FDSC introduces a multi-scale pooling feature fusion (MSPF2) module. Finally, to better focus on important features and reduce the attention to unimportant or redundant information, RVS-FDSC introduces a residual parallel dual attention (RPDA) module. We conduct experiments on three retinal vessel segmentation datasets: DRIVE, CHASE-DB1, and STARE. RVS-FDSC achieves excellent results on these three public datasets, with AUC of 0.9856, 0.9867, and 0.9808, Acc of 0.9692, 0.9743, and 0.9718, and Spe of 0.9878, 0.9856 and 0.9882, respectively. These results indicate the superior performance of RVS-FDSC compared to other state-of-the-art methods.

A novel Fourier neural operator framework for classification of multi-sized images: Application to three dimensional digital porous media

Fourier neural operators (FNOs) are invariant with respect to the size of input images, and thus images with any size can be fed into FNO-based frameworks without any modification of network architectures, in contrast to traditional convolutional neural networks. Leveraging the advantage of FNOs, we propose a novel deep-learning framework for classifying images with varying sizes. Particularly, we simultaneously train the proposed network on multi-sized images. As a practical application, we consider the problem of predicting the label (e.g., permeability) of three-dimensional digital porous media. To construct the framework, an intuitive approach is to connect FNO layers to a classifier using adaptive max pooling. First, we show that this approach is only effective for porous media with fixed sizes, whereas it fails for porous media of varying sizes. To overcome this limitation, we introduce our approach: instead of using adaptive max pooling, we use static max pooling with the size of channel width of FNO layers. Since the channel width of the FNO layers is independent of the input image size, the introduced framework can handle multi-sized images during training. We show the effectiveness of the introduced framework and compare its performance with the intuitive approach through the example of the classification of three-dimensional digital porous media of varying sizes.

Quantum driven deep learning for enhanced diabetic retinopathy detection

Traditional convolutional neural networks (CNNs) have shown potential for recognizing retinopathy caused by diabetes (DR). However, developing quantum computing has the possibility for improved feature representation. We propose a hybrid approach that combines classical CNNs with quantum circuits to capitalize on both classical and quantum information for DR classification. Using the Keras and Qiskit frameworks, our model encodes picture features into quantum states, allowing for richer representations. Through experiments on a collection of retinal pictures, our model displays competitive performance, with excellent reliability and precision in categorizing DR severity levels. This combination of classical and quantum paradigms offers a fresh approach to enhancing DR diagnosis and therapy.

Enhanced Vision Transformer and Transfer Learning Approach to Improve Rice Disease Recognition

In the evolving landscape of agricultural technology, recognizing rice diseases through computational models is a critical challenge, predominantly addressed through Convolutional Neural Networks (CNN). However, the localized feature extraction of CNNs often falls short in complex scenarios, necessitating a shift towards models capable of global contextual understanding. Enter the Vision Transformer (ViT), a paradigm-shifting deep learning model that leverages a self-attention mechanism to transcend the limitations of CNNs by capturing image features in a comprehensive global context. This research embarks on an ambitious journey to refine and adapt the ViT Base(B) transfer learning model for the nuanced task of rice disease recognition. Through meticulous reconfiguration, layer augmentation, and hyperparameter tuning, the study tests the model's prowess across both balanced and imbalanced datasets, revealing its remarkable ability to outperform traditional CNN models, including VGG, MobileNet, and EfficientNet. The proposed ViT model not only achieved superior recall (0.9792), precision (0.9815), specificity (0.9938), f1-score (0.9791), and accuracy (0.9792) on challenging datasets but also established a new benchmark in rice disease recognition, underscoring its potential as a transformative tool in the agricultural domain. This work not only showcases the ViT model's superior performance and stability across diverse tasks and datasets but also illuminates its potential to revolutionize rice disease recognition, setting the stage for future explorations in agricultural AI applications.

LSGNet: A lightweight convolutional neural network model for tomato disease identification

Tomatoes are among the most extensively grown and consumed crops worldwide, but tomato production can be greatly reduced due to various diseases. Plant diseases show different symptoms at different stages. In addition, there are similarities in the symptoms of different types of plant diseases, which hinder the recognition of diseases by existing deep learning models. Traditional convolutional neural network (CNN) models for disease recognition have a large number of parameters that require high computational resources. To overcome these challenges, we propose a lightweight CNN model named LSGNet (lightweight sandglass network) for tomato disease identification. The LSGNet backbone consists of the sandglass with efficient channel attention (SGECA) and the position aware circular convolution sandglass (ParcSG) modules. The SGECA module reduces the interference of complex environments and thus focuses on extracting useful feature information. The ParcSG module has a global receptive field, which provides more detailed feature information on disease recognition. The results show that the recognition accuracies are 92.37%, 94.32%, 89.64%, 92.70%, 94.43%, 90.97%, 89.42%, 92.98%, 89.58%, and 95.54% for AlexNet, ResNet50, VGG16, MobileNetV3-Large, ShuffleNetV2-1 × , EfficientNetV2-Small, ViT-Base, MobileViT-Small, Swin-Tiny, and LSGNet. Therefore, LSGNet has higher accuracy in recognizing tomato diseases compared to other classical models. In addition, LSGNet uses 0.75 million parameters. Compared to the lightweight CNN model MobileNetV3-Large, it only has 18% of the parameters. As a whole, the advantages of LSGNet in efficiency and lightweight structure would make it a useful tool for tomato disease recognition on mobile or embedded devices.

DenseHillNet: a lightweight CNN for accurate classification of natural images.

The detection of natural images, such as glaciers and mountains, holds practical applications in transportation automation and outdoor activities. Convolutional neural networks (CNNs) have been widely employed for image recognition and classification tasks. While previous studies have focused on fruits, land sliding, and medical images, there is a need for further research on the detection of natural images, particularly glaciers and mountains. To address the limitations of traditional CNNs, such as vanishing gradients and the need for many layers, the proposed work introduces a novel model called DenseHillNet. The model utilizes a DenseHillNet architecture, a type of CNN with densely connected layers, to accurately classify images as glaciers or mountains. The model contributes to the development of automation technologies in transportation and outdoor activities. The dataset used in this study comprises 3,096 images of each of the "glacier" and "mountain" categories. Rigorous methodology was employed for dataset preparation and model training, ensuring the validity of the results. A comparison with a previous work revealed that the proposed DenseHillNet model, trained on both glacier and mountain images, achieved higher accuracy (86%) compared to a CNN model that only utilized glacier images (72%). Researchers and graduate students are the audience of our article.

A Performance Comparison of Japanese Sign Language Recognition with ViT and CNN Using Angular Features

In recent years, developments in deep learning technology have driven significant advancements in research aimed at facilitating communication with individuals who have hearing impairments. The focus has been on enhancing automatic recognition and translation systems for sign language. This study proposes a novel approach using a vision transformer (ViT) for recognizing Japanese Sign Language. Our method employs a pose estimation library, MediaPipe, to extract the positional coordinates of each finger joint within video frames and generate one-dimensional angular feature data from these coordinates. Then, the code arranges these feature data in a temporal sequence to form a two-dimensional input vector for the ViT model. To determine the optimal configuration, this study evaluated recognition accuracy by manipulating the number of encoder layers within the ViT model and compared against traditional convolutional neural network (CNN) models to evaluate its effectiveness. The experimental results showed 99.7% accuracy for the method using the ViT model and 99.3% for the results using the CNN. We demonstrated the efficacy of our approach through real-time recognition experiments using Japanese sign language videos.

An Effective Image Classification Method for Plant Diseases with Improved Channel Attention Mechanism aECAnet Based on Deep Learning

Since plant diseases occurring during the growth process are a significant factor leading to the decline in both yield and quality, the classification and detection of plant leaf diseases, followed by timely prevention and control measures, are crucial for safeguarding plant productivity and quality. As the traditional convolutional neural network structure cannot effectively recognize similar plant leaf diseases, in order to more accurately identify the diseases on plant leaves, this paper proposes an effective plant disease image recognition method aECA-ResNet34. This method is based on ResNet34, and in the first and the last layers of this network, respectively, we add this paper’s improved aECAnet with the symmetric structure. aECA-ResNet34 is compared with different plant disease classification models on the peanut dataset constructed in this paper and the open-source PlantVillage dataset. The experimental results show that the aECA-ResNet34 model proposed in this paper has higher accuracy, better performance, and better robustness. The results show that the aECA-ResNet34 model proposed in this paper is able to recognize diseases of multiple plant leaves very accurately.