Abstract: A modified fuzzy c-means clustering algorithm based on HSI color space was proposed and introduced for the separation of fascia region in color beef ribeye image. In this method, the H component and I component were used as feature vectors according to the distribution characteristics of beef ribeye image pixels to modify initial cluster center selection and distance-weight formula for FCM clustering algorithm. The results obtained showed that the proposed HSI algorithm could ensure fast convergence features of FCM clustering and increase the precision and accuracy of image segmentation, and had strong noise-suppressing capability.

Key words: fuzzy c-means clustering, image segmentation, HSI model, beef image, fascia recognition.