LFRE-YOLO: Lightweight Edge Computing Algorithm for Detecting External-Damage Objects on Transmission Lines

Convolutional neural networks (CNNs) have achieved significant accuracy improvement in many intelligent applications at the cost of intensive convolution operations and massive data movements. To efficiently deploy CNNs on low power embedded platforms in real time, the depthwise separable convolution has been proposed to replace the standard convolution, especially in lightweight CNNs, which remarkably reduces the computation complexity and model size. However, it is difficult for a general convolution engine to obtain the theoretical performance improvement as the decreased data dependency of depthwise convolution significantly reduces the data reuse opportunity. To address this issue, a flexible and highperformance accelerator based on FPGA is proposed to efficiently process the inference of both large-scale and lightweight CNNs. Firstly, by sharing the activation dataflow between the depthwise convolution and pooling layers, the control logic and data bus of the two layers are reused to maximize the data utilization and minimize the logic overhead. Furthermore, these two layers can be processed either directly after standard convolutions to eliminate the external memory accesses or independently to gain better flexibility. Thirdly, a performance model is proposed to automatically explore the optimal design options of the accelerator. The proposed hardware accelerator is evaluated on Intel Arria 10 SoC FPGA and demonstrates state-of-the-art performance on both large-scale CNNs, e.g., VGG, and lightweight ones, e.g., MobileNet.

Convolutional neural networks (CNNs) are becoming increasingly popular in modern remote sensing image processing tasks and exhibit outstanding capability for hyperspectral image (HSI) classification. However, for the existing CNN-based HSI-classification methods, most of them only consider single-scale feature extraction, which may neglect some important fine information and cannot guarantee to capture optimal spatial features. Moreover, many state-of-the-art methods have a huge number of network parameters needed to be tuned, which will cause high computational cost. To address the aforementioned two issues, a novel multiscale residual network (MSRN) is proposed for HSI classification. Specifically, the proposed MSRN introduces depthwise separable convolution (DSC) and replaces the ordinary depthwise convolution in DSC with mixed depthwise convolution (MDConv), which mixes up multiple kernel sizes in a single depthwise convolution operation. The DSC with mixed depthwise convolution (MDSConv) can not only explore features at different scales from each feature map but also greatly reduce learnable parameters in the network. In addition, a multiscale residual block (MRB) is designed by replacing the convolutional layer in an ordinary residual block with the MDSConv layer. The MRB is used as the major unit of the proposed MSRN. Furthermore, to enhance further the feature representation ability, the proposed network adds a high-level shortcut connection (HSC) on the cascaded two MRBs to aggregate lower level features and higher level features. Experimental results on three benchmark HSIs demonstrate the superiority of the proposed MSRN method over several state-of-the-art methods.

Bird activities seriously affect the safety of power lines. An improved lightweight YOLOv4 model was proposed to recognize typical bird species threatening power transmission line security in this study. A dataset composed of 3000 images about 10 bird species that easily cause transmission line outages was constructed. An improved YOLOv4 model was established by replacing the feature extraction network with GhostNet. The focus layer was added in GhostNet, and the standard convolution in the path aggregation network (PANet) was replaced with the depthwise separable convolution (DSC), thus to greatly reduce the number of parameters in the model. After model training, the improved YOLOv4 was applied to detect bird targets in 300 test sample images. The experimental results indicate that the mean average precision (mAP) and frames per second (FPS) of the proposed model are respectively 97.55 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> and 43, which is much faster than YOLOv4. In terms of detection accuracy and efficiency, the proposed model was compared to the existing models such as SSD and YOLOv4. This study can be applied for bird recognition and therefore contribute to achieve differentiated prevention of bird-related outages.

One difficult problem of keyword spotting is how to miniaturize its memory footprint while maintain a high precision. Although convolutional neural networks have shown to be effective to the small-footprint keyword spotting problem, they still need hundreds of thousands of parameters to achieve good performance. In this paper, we propose an efficient model based on depthwise separable convolution layers and squeeze-and-excitation blocks. Specifically, we replace the standard convolution by the depthwise separable convolution, which reduces the number of the parameters of the standard convolution without significant performance degradation. We further improve the performance of the depthwise separable convolution by reweighting the output feature maps of the first convolution layer with a so-called squeeze-and-excitation block. We compared the proposed method with five representative models on two experimental settings of the Google Speech Commands dataset. Experimental results show that the proposed method achieves the state-of-the-art performance. For example, it achieves a classification error rate of 3.29% with a number of parameters of 72K in the first experiment, which significantly outperforms the comparison methods given a similar model size. It achieves an error rate of 3.97% with a number of parameters of 10K, which is also slightly better than the state-of-the-art comparison method given a similar model size.

With the rapid growth of artificial intelligence applications on embedded devices and mobile platforms, traditional convolutional neural network models are often difficult to meet the real-time and power consumption requirements of embedded devices due to their large number of parameters. This paper presents a lightweight improvement to the GoogLeNet model by replacing the standard convolution in the Inception module with Depthwise Separable Convolutions and simplifying the network structure. This approach effectively reduces the model’s parameter count and computational complexity. Experimental results show that the improved model reduces the parameter count by approximately $54 \%$ and the computational complexity by about $42 \%$. The training speed is significantly improved, disturbance resistance is enhanced, and the model is suitable for deployment on low-computational-power devices.

Intelligent fault diagnosis methods based on Transformer variants have shown excellent performance. However, these methods generally model features globally, which means a poor anti-noise ability and greater computational complexity. Therefore, a lightweight framework for bearing fault diagnosis based on depthwise separable convolution (DSC) and sparse attention mechanism named DSC-SparseFormer is proposed. First, a guided-and-sparse self-attention (GSSA) module, which can adaptively adjust the region of interest for features, is proposed to extract features, thereby reducing sensitivity to redundant information and downgrading computational complexity. This is potentially the first time such a methodology has been proposed and applied. Second, discrete cosine transform (DCT) is applied to map time-domain signals into frequency domain to enrich feature space, and a time-frequency coupling attention (TFCA) module is developed to dynamically integrate time-domain and frequency-domain features. Finally, extensive experiments on the HUSTBearing dataset, DHU dataset, and HIT dataset demonstrate that DSC-SparseFormer achieves high-accuracy (HUSTBearing: 99.67%, DHU: 99.94%, and HIT: 99.93%) and lightweight (Params: 0.003 M, FLOPs: 6.07 M) fault diagnosis. The effectiveness of the GSSA and TFCA is validated by ablation studies, and feature visualisation is employed to improve the interpretability of the DSC-SparseFormer.

Aimed to reduce the excessive cost of neural network, this paper proposes a lightweight neural network combining dilated convolution and depthwise separable convolution. Firstly, the dilated convolution is used to expand the receptive field during the convolution process while maintaining the number of convolution parameters, which can extract more high-level global semantic features and improve the classification accuracy of the network. Second, the use of the depthwise separable convolution reduces the network parameters and computational complexity in convolution operations. The experimental results on the CIFAR-10 dataset show that the proposed method improves the classification accuracy of the network while effectively compressing the network size.KeywordsLightweight neural networkDilated convolutionDepthwise separable convolutionClassification accuracyCloud computing

Modern efficient Convolutional Neural Networks(CNNs) always use Depthwise Separable Convolutions(DSCs) and Neural Architecture Search(NAS) to reduce the number of parameters and the computational complexity. But some inherent characteristics of networks are overlooked. Inspired by visualizing feature maps and N×N(N>1) convolution kernels, several guidelines are introduced in this paper to further improve parameter efficiency and inference speed. Based on these guidelines, our parameter-efficient CNN architecture, called VGNetG, achieves better accuracy and lower latency than previous networks with about 30%~50% parameters reduction. Our VGNetG-1.0MP achieves 67.7% top-1 accuracy with 0.99M parameters and 69.2% top-1 accuracy with 1.14M parameters on ImageNet classification dataset. Furthermore, we demonstrate that edge detectors can replace learnable depthwise convolution layers to mix features by replacing the N×N kernels with fixed edge detection ker-nels. And our VGNetF-1.5MP archives 64.4%(-3.2%) top-1 accuracy and 66.2%(-1.4%) top-1 accuracy with additional Gaussian kernels.

Diseases such as diabetes and hypertension can lead to change the shape of the retinal blood vessels. Segmentation of fundus images is a key step in the process of quantitative analysis of the disease, which is instructive in the analysis and diagnosis of clinical diseases. In this paper, a method for the segmentation of retinal image vessels based on fully convolutional network (FCN) with depthwise separable convolution and channel weighting is presented. Firstly, CLAHE and Gamma correction of the green channel of the fundus image are used to enhance the contrast. Then, in order to adapt to network training, the enhanced image is divided into patches to expand the data. Finally, the depthwise separable convolution instead of the standard convolution method is used to increase the network width. Meanwhile, the channel weighting module is introduced to explicitly model the relationship between the characteristic channels in order to improve the distinguishability of the features. The combination of them is applied to the FCN and the results of expert manual identification are used to supervise the experiment on the DRIVE database. The results show that the segmentation accuracy of the proposed method in DRIVE database reached 0.963 0 and AUC reached 0.983 1. The segmentation accuracy in STARE database reached 0.962 0 and AUC achieved 0.983 0. To some extent, the proposed method has better feature resolution and better segmentation performance.

To address the issues of excessive computational complexity and inadequate feature extraction capability in traditional motor bearing fault detection models, this study proposes an improved DSC-EMA-VGG16 algorithm based on the VGG16 architecture. By optimizing the structure of the VGG16 algorithm and integrating novel modules, the algorithm achieves a reduction in parameter count while enhancing feature extraction performance. Firstly, a DSC-EMA module is designed, which incorporates the EMA (Efficient Multi-Scale Attention) mechanism and replaces partial standard convolution modules within the EMA module with the DSC (Depthwise Separable Convolution) module. This DSC-EMA module is embedded into the forward propagation layer of the VGG16 network, leading to a significant reduction in the model's computational complexity and parameter volume, an improvement in its multi-scale feature fusion capability, and a subsequent enhancement in feature extraction efficiency. Secondly, the LeakyReLU activation function is adopted to replace the ReLU activation function in the original VGG16 network, further boosting the model's training efficiency and feature extraction capacity. Experimental results demonstrate that the DSC-EMA-VGG16 algorithm exhibits outstanding performance in motor bearing fault detection tasks, achieving an accuracy of 99.1%. This fully meets the requirements for feature extraction in motor bearing fault detection and provides a novel approach for addressing complex motor bearing fault scenarios.

There is a need for transmission and distribution to supply high-quality, dependable electricity to customers. The detection of transmission line problems must be more precise, consistent, and thorough in order to achieve acceptable power quality. The first part of the thesis focuses on HV transmission line trip rates and operational defects during the last several years. As shown by the recordings, there are a variety of causes for electrical failures and tripping, including lightning strikes, wind deviation discharge, external force damage, bird catastrophes, covered ice and waving, pollution flashovers, undetermined explanations, and more. The quality of power is affected by problems with transmission and distribution lines. Circuit breakers reclose automatically when a fault occurs because the control circuit is synchronized with the breaker. There is no need for human intervention. When a transitory problem develops at the distribution level of the system, the auto recloser helps to keep the system up and running for a longer period of time. This paper demonstrates a unique structured interconnected structure for fault level and classification approaches for transmission lines, as well as a method for overcoming this technique based on the autoreclosing characteristic of the circuit breaker.

PBR-YOLO: A lightweight piglet multi-behavior recognition algorithm based on improved yolov8

Hybrid deep learning architecture for skin disease classification

Addressing the challenges of multi-scene precision pesticide application for field broccoli crops and computational limitations of edge devices, this study proposes a lightweight broccoli detection method named DWG-YOLOv8, based on an improved YOLOv8s architecture. Firstly, Ghost Convolution is introduced into the C2f module, and the standard CBS module is replaced with Depthwise Separable Convolution (DWConv) to reduce model parameters and computational load during feature extraction. Secondly, a CDSL module is designed to enhance the model’s feature extraction capability. The CBAM attention mechanism is incorporated into the Neck network to strengthen the extraction of channel and spatial features, enhancing the model’s focus on the target. Experimental results indicate that compared to the original YOLOv8s, the DWG-YOLOv8 model has a size decreased by 35.6%, a processing time reduced by 1.9 ms, while its precision, recall, and mean Average Precision (mAP) have increased by 1.9%, 0.9%, and 3.4%, respectively. In comparative tests on complex background images, DWG-YOLOv8 showed reductions of 1.4% and 16.6% in miss rate and false positive rate compared to YOLOv8s. Deployed on edge devices using field-collected data, the DWG-YOLOv8 model achieved a comprehensive recognition accuracy of 96.53%, representing a 5.6% improvement over YOLOv8s. DWG-YOLOv8 effectively meets the lightweight requirements for accurate broccoli recognition in complex field backgrounds, providing technical support for object detection in intelligent precision pesticide application processes for broccoli.

In order to improve the perception accuracy of the apple tree fruit recognition model and to reduce the model size, a lightweight apple target recognition method based on an improved YOLOv5s artificial intelligence algorithm was proposed, and relevant experiments were designed. The Depthwise Separable Convolution (DWConv) module has many advantages: (1) It has high computational efficiency, reducing the number of parameters and calculations in the model; (2) It makes the model lightweight and easy to deploy in hardware; (3) DWConv can be combined with other modules to enhance the multi-scale feature extraction capability of the detection network and improve the ability to capture multi-scale information; (4) It balances the detection accuracy and speed of the model; (5) DWConv can flexibly adapt to different network structures. Because of its efficient computing modes, lightweight design, and flexible structural adaptation, the DWConv module has significant advantages in multi-scale feature extraction, real-time performance improvement, and small-object detection. Therefore, this method improves the original YOLOv5s network architecture by replacing the embedded Depthwise Separable Convolution in its Backbone network, which reduces the size and parameter count of the model while ensuring detection accuracy. The experimental results show that for the test-set images, the proposed improved model has an average recognition accuracy of 92.3% for apple targets, a recognition time of 0.033 s for a single image, and a model volume of 11.1 MB. Compared with the original YOLOv5s model, the average recognition accuracy was increased by 0.8%, the recognition speed was increased by 23.3%, and the model volume was compressed by 20.7%, effectively achieving lightweight improvement of the apple detection model and improving the accuracy and speed of detection. The detection algorithm proposed in the study can be extended to the intelligent measurement of apple biological and physical characteristics, including for size measurement, shape analysis, and color analysis. The proposed method can improve the intelligence level of orchard management and horticultural technology, reduce labor costs, assist precision agriculture technology, and promote the transformation of the horticultural industry toward sustainable development.