Journal of Digital Information Management Hongwei YUE, Ken CAI, Bing LUO, Yingying JIN, Zhaofeng ZENG 13 1 2015 https://www.dline.info/fpaper/jdim/v13i1/v13i1_4.pdf In this study,mean shift algorithm and region merging were combined to automatically segment a digital feather image and remove the noise in digital images more effectively for segmentation of a feather quill and a feather leaf. First, the mean shift algorithm employed to calculate the convergence value of each pixel can obtain a filtered smoothened image; then, setting region merging criteria were used to merge filtered images. Finally, the threshold segmentation algorithm was used to extract the feather quill. Experimental results indicated that the mean shift algorithm could be used to denoise the digital feather images, thereby achieving a high value of evaluation index and improved visual quality.Moreover, this algorithm can accurately segment texture image and effectively filter out unwanted background texture information to segment texture image with unclear color information.