In computer vision technology, image segmentation is a significant technological advancement for the current problems of high-speed railroad image scene changes, low segmentation accuracy, and serious information loss. We propose a segmentation algorithm, DFA-UNet, based on an improved U-Net network architecture. The model uses the same encoder–decoder structure as U-Net. To be able to extract image features efficiently and further integrate the weights of each channel feature, we propose to embed the DFA attention module in the encoder part of the model for the adaptive adjustment of feature map weights. We evaluated the performance of the model on the RailSem19 dataset. The results showed that our model showed improvements of 2.48%, 0.22%, 3.31%, 0.97%, and 2.2% in mIoU, F1-score, Accuracy, Precision, and Recall, respectively, compared with U-Net. The model can effectively achieve the segmentation of railroad images.

Retinal vessel segmentation remains a challenging task because the morphology of the retinal vessels reflects the health of a person, which is essential for clinical diagnosis. Therefore, achieving accurate segmentation of the retinal vessel shape can determine the patient’s physical condition in a timely manner and can prevent blindness in patients. Since the traditional retinal vascular segmentation method is manually operated, this can be time-consuming and laborious. With the development of convolutional neural networks, U-shaped networks (U-Nets) and variants show good performance in image segmentation. However, U-Net is prone to feature loss due to the operation of the encoder convolution layer and also causes the problem of mismatch in the processing of contextual information features caused by the skip connection part. Therefore, we propose an improvement of the retinal vessel segmentation method based on U-Net to segment retinal vessels accurately. In order to extract more features from encoder features, we replace the convolutional layer with ResNest network structure in feature extraction, which aims to enhance image feature extraction. In addition, a Depthwise FCA Block (DFB) module is proposed to deal with the mismatched processing of local contextual features by skip connections. Combined with the two public datasets on retinal vessel segmentation, namely DRIVE and CHASE_DB1, and comparing our method with a larger number of networks, the experimental results confirmed the effectiveness of the proposed method. Our method is better than most segmentation networks, demonstrating the method’s significant clinical value.

As artificial intelligence develops, deep learning algorithms are increasingly being used in the field of dorsal hand vein (DHV) recognition. However, deep learning has high requirements regarding the number of samples, and current DHV datasets have few images. To solve the above problems, we propose a method based on the fusion of ResNet and Histograms of Oriented Gradients (HOG) features, in which the shallow semantic information extracted by primary convolution and HOG features are fed into the residual structure of ResNet for full fusion and, finally, classification. By adding Gaussian noise, the North China University of Technology dataset, the Shandong University of Science and Technology dataset, and the Eastern Mediterranean University dataset are extended and fused to from a fused dataset. Our proposed method is applied to the above datasets, and the experimental results show that our proposed method achieves good recognition rates on each of the datasets. Importantly, we achieved a 93.47% recognition rate on the fused dataset, which was 2.31% and 26.08% higher than using ResNet and HOG alone.

In the context of COVID-19, the research on various aspects of the venipuncture robot field has become increasingly hot, but there has been little research on robotic needle insertion angles, primarily performed at a rough angle. This will increase the rate of puncture failure. Furthermore, there is sometimes significant pain due to the patients’ differences. This paper investigates the optimal needle entry angle decision for a dorsal hand intravenous injection robot. The dorsal plane of the hand was obtained by a linear structured light scan, which was used as a basis for calculating the needle entry angle. Simulation experiments were also designed to determine the optimal needle entry angle. Firstly, the linear structured optical system was calibrated and optimized, and the error function was constructed and solved iteratively by the optimization method to eliminate measurement error. Besides, the dorsal hand was scanned to obtain the spatial point clouds of the needle entry area, and the least squares method was used to fit it to obtain the dorsal hand plane. Then, the needle entry angle was calculated based on the needle entry area plane. Finally, the changes in the penetration force under different needle entry angles were analyzed to determine the optimal needle insertion angle. According to the experimental results, the average error of the optimized structured light plane position was about 0.1 mm, which meets the needs of the project, and a large angle should be properly selected for needle insertion during the intravenous injection.