Image Similarity Check of Nailfold Capillary by Template Matching

Sayaka Suga, Aoi Otomo, Hieyong Jeong, Yuko Ohno

Abstract: Objective: This study was to propose a continuous observation method of same nailfold capillaries for the early diagnosis of secondary Raynauds phenomenon through the image processing. Methods: After checking the amount of blurring, a panorama image, which enabled us to understand the whole condition of nailfold capillaries and decide the target area for the continuous observation, was made by captured images with the microscope. After the targeted area was automatically detected, the image similarity check between the previous and the current captured capillary images was evaluated in the real-time through the template matching. Simultaneously, the edge detection allowed us to evaluate the targeted area with coordinates of capillaries as quantitative method. Discussion: Because the observation of continuous same nailfold capillaries enabled us to help the early diagnosis of systemic sclerosis, it was discussed that the low-cost proposed method based on the software approach had the clinical efficacy in rheumatic diseases.