Myoelectric prosthetic hands are controlled by surface electromyographic signals from residual muscles in amputees. Recognizing continuous finger movements is crucial for improving dexterity when switching tasks. However, existing finger motion decoding algorithms are limited to recognizing discrete static gestures and are unable to cope with the continuous state switching of fine finger movements. To overcome the challenge, a hybrid fuzzy temporal transformer network is proposed. The network integrates temporal convolutional networks with transformers to extract temporal features of motor unit action potentials. It addresses the challenge of identifying waveform changes during gesture switching. An integrated fuzzy decision layer dynamically adjusts the recognition thresholds for static and switching states to suppress noise interference caused by muscle fatigue. By analyzing the impact of switching duration on waveform distortion, the network achieves duration detection for finger movement switching, enabling prediction of target gestures to enhance operational continuity and dexterity. The experiment involved 10 subjects switching between their thumb, index finger, and little finger for different durations. The results show that the correlation between finger switching recognition and gesture targets is 0.859, which is at least 0.026 above that of the baseline methods, and an F1-score at least 0.069 higher. The detection error for switching duration is reduced to a minimum of 80 ms. This work provides a feasible solution for controlling dexterous prostheses in temporal movement switching tasks. Through the collaborative design of hybrid architecture, it paves the way for novel solutions in the practical application of myoelectric prosthetic hand dexterous control.

Image compression is increasingly important in applications like intelligent driving and smart surveillance systems. This study presents a novel cross view capture distributed image compression network (CVCDIC) to improve the compression quality by using decoder side information. The CVCDIC’s decoder utilizes feature extraction networks to extract features from both the primary image and the side information. Furthermore, a multi-level cross view attention module is designed to capture interrelated details between images at multiple hierarchical levels. Finally, a spatial refinement module, constructed on the foundation of information distillation networks, is designed to further refine the quality of reconstructed images. The results show that CVCDIC can achieve an MS-SSIM of 0.978 at 0.15 bpp, surpassing DSIN (0.925), NDIC (0.956), and ATN (0.955) on the KITTI Stereo dataset.