Tsai C.-TJIAN-JIUN DING2022-04-252022-04-252021https://www.scopus.com/inward/record.uri?eid=2-s2.0-85124154538&doi=10.1109%2fISPACS51563.2021.9651060&partnerID=40&md5=b6c948da9d09a8297ebacd2354e50f12https://scholars.lib.ntu.edu.tw/handle/123456789/607201In this paper, an automatic algorithm for baseball rotation analysis. To analyze the rotation angle of a baseball, the seam feature and its curvature play very important roles. Therefore, after performing background removal and denoising, we first adopt ridge detection to extract the seam on the baseball. Then, the curvature is calculated. According to the pixel-wise curvature information, one can determine the ration angle based on template sifting and matching. With the proposed algorithm, the rotation angle of the baseball can be determined accurately. It is helpful for determining the rotation rate of the baseball and beneficial for automatic sport video analysis. ? 2021 IEEE.Background subtractionCurvatureRidge detectionSeam matchingSport videoPixelsRotation rateIdentification algorithmsMatchingsPixel similaritiesRidge detectionsRotation anglesSeam curvatureBaseball[SDGs]SDG14conference paper10.1109/ISPACS51563.2021.96510602-s2.0-85124154538