Multi_scale_precise_registration_and_anti-shake_algorithm_for_nail_fold_vascular_area_perception

In nailfold capillary imaging, the inevitable physiological shaking of the human hand is drastically magnified from the microscopic perspective, causing image shaking and interfering with the accurate measurement of parameters such as capillary diameter and blood flow velocity. Thus, registration algorithms must be used for image stabilization and suppressing shaking. Moreover, the nonplanar structure of nailfolds provides a distribution of capillary bundles at different depths. When the depth difference exceeds the camera depth of field, large-field nailfold capillary images are prone to missing vascular information and require zooming to obtain global vascular information. Although using a large-depth-of-field camera can alleviate this problem, it is has a high cost, and the lack of vascular information hinders nailfold capillary video registration. As a comprehensive solution, we combine the YOLOv8 detection model with the Laplacian clarity evaluation function for real-time tracking of focus changes during capillary imaging. In addition, normalized cross-correlation template matching is used to calculate the optimal matching position of adjacent frames, determine the position with the highest vascular similarity, and obtain the optimal offset between frames for accurate registration. Experimental results show the high performance of registered videos for three evaluation indicators: mean squared error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM). Compared with the original video, MSE decreases by 64.5%, PSNR increases by 3.88 dB, and SSIM increases by 4.7% in zooming experiments. Moreover, MSE decreases by 57.9%, PSNR increases by 3.99 dB, and SSIM increases by 3.19% in fixed-focus experiments. Compared with state-of-the-art registration algorithms, our proposal performs better in overall performance across the three evaluation indicators, achieving the best registration effect for capillary images.