Wireless capsule endoscope (WCE) has been verified in clinical medicine for many years. However, the detection process needs experienced doctors to read the film manually for a long time. In addition, the cost of the endoscope itself leads to a high cost of WCE detection and overall cycle is long. New research method based on deep learning technology with robustness and high accuracy can reduce the detection cost and benefit the public. According to the characteristics of small intestine lesion, a method focuses on labeling and feature detection which can optimize the process by analyzing small intestine WCE image and experimental comparison. Based on the YOLOv3 detection network, retaining the original basic feature of detection network, an improved one is further optimized and effectively verified. Finally, the redundant images are filtered out by comparing the Hash value of images, presenting the final concise detection results for doctors. Starting from image labeling, the design of deep learning network structure for the image of small intestine digestive tract endoscope is studied, which can effectively improve intelligent detection computer-aided clinical application of WCE, with higher accuracy and lower missing detection rate than manual detection.
Convolutional neural networks (CNNs) have made significant progress in the field of facial expression recognition (FER). However, due to challenges such as occlusion, lighting variations, and changes in head pose, facial expression recognition in real-world environments remains highly challenging. At the same time, methods solely based on CNN heavily rely on local spatial features, lack global information, and struggle to balance the relationship between computational complexity and recognition accuracy. Consequently, the CNN-based models still fall short in their ability to address FER adequately. To address these issues, we propose a lightweight facial expression recognition method based on a hybrid vision transformer. This method captures multi-scale facial features through an improved attention module, achieving richer feature integration, enhancing the network’s perception of key facial expression regions, and improving feature extraction capabilities. Additionally, to further enhance the model’s performance, we have designed the patch dropping (PD) module. This module aims to emulate the attention allocation mechanism of the human visual system for local features, guiding the network to focus on the most discriminative features, reducing the influence of irrelevant features, and intuitively lowering computational costs. Extensive experiments demonstrate that our approach significantly outperforms other methods, achieving an accuracy of 86.51% on RAF-DB and nearly 70% on FER2013, with a model size of only 3.64 MB. These results demonstrate that our method provides a new perspective for the field of facial expression recognition.
Image semantic segmentation represents a significant area of research within the field of computer vision. With the advent of deep learning, image semantic segmentation techniques that integrate deep learning have demonstrated superior accuracy compared to traditional image semantic segmentation methods. Recently, the Mamba architecture has demonstrated superior semantic segmentation performance compared to the Transformer architecture, and has consequently become a research focus in this field. Nevertheless, the specifics of the Mamba architecture have remained underexplored in the extant literature. This review provides a comprehensive overview of the latest research progress in deep learning techniques for semantic segmentation. It offers a systematic review of traditional convolutional neural network (CNN)-based architectures and focuses on a series of emerging architectures, including the Transformer architecture, the Mamba architecture, and cutting-edge approaches such as self-supervised learning strategies. For each category, a detailed account is provided of the principal algorithms and techniques employed, together with a report on the performance achieved using datasets commonly used in the field.
In recent years, image semantic segmentation algorithms have made significant progress driven by deep learning technology, and are widely used in fields such as medical image analysis, assistive technology for the blind, and autonomous driving. Aiming at problems such as the inability of many image segmentation algorithms to fully capture global context information, low computational efficiency, and insufficient context information fusion. This article integrates the Transformer mechanism and CA mechanism into the DeepLabV3+ network and proposes the TransDeep network. In the encoder, two different backbones, Xception and MobileNetV2, are first used for feature extraction, and a better backbone network is selected. Furthermore, based on the lightweight backbone network, the Transformer mechanism is integrated into the advanced features of the backbone to enhance long-range dependence. Secondly, the Coord Attention module (CA) is added after the low-level features of the backbone to strengthen information such as edge and detail features. Finally, the Coord Attention mechanism (CA) is added after the ASPP module to allow the model to focus on key image features while effectively filtering out irrelevant background information. Experimental results show that the TransDeep network can improve the accuracy of key categories and effectively improve the network’s segmentation accuracy of targets in images. It achieved an MIoU of 73.5% on the Pascal test set and a good performance of 79.95% on the CamVid test set. Compared with the baseline model, they improved by 2.10% and 2.61% respectively.
To address the issues of inadequate feature extraction for rolling bearings, inaccurate fault diagnosis, and overfitting in complex operating conditions, this paper proposes a rolling bearing diagnosis method based on multi-scale feature fusion and transfer adversarial learning. Firstly, a multi-scale convolutional fusion layer is designed to effectively extract fault features from the original vibration signals at multiple time scales. Through a feature encoding fusion module based on the multi-head attention mechanism, feature fusion extraction is performed, which can model long-distance contextual information and significantly improve diagnostic accuracy and anti-noise capability. Secondly, based on the domain adaptation (DA) cross-domain feature adversarial learning strategy of transfer learning methods, the extraction of optimal domain-invariant features is achieved by reducing the gap in data distribution between the target domain and the source domain, addressing the call for research on fault diagnosis across operating conditions, equipment, and virtual–real migrations. Finally, experiments were conducted to verify and optimize the effectiveness of the feature extraction and fusion network. A public bearing dataset was used as the source domain data, and special vehicle bearing data were selected as the target domain data for comparative experiments on the effect of network transfer learning. The experimental results demonstrate that the proposed method exhibits an exceptional performance in cross-domain and variable load environments. In multiple bearing cross-domain transfer learning tasks, the method achieves an average migration fault diagnosis accuracy rate of up to 98.65%. When compared with existing methods, the proposed method significantly enhances the ability of data feature extraction, thereby achieving a more robust diagnostic performance.