作者: H.D. Cheng , X.H. Jiang , Jingli Wang
DOI: 10.1016/S0031-3203(01)00054-1
关键词: Computer vision 、 Color normalization 、 Region growing 、 Computer science 、 Histogram 、 Thresholding 、 Color space 、 Image segmentation 、 Artificial intelligence 、 Color histogram 、 Pixel
摘要: In this paper, a color image segmentation approach based on homogram thresholding and region merging is presented. The considers both the occurrence of gray levels neighboring homogeneity value among pixels. Therefore, it employs local global information. Fuzzy entropy utilized as tool to perform analysis for finding all major homogeneous regions at first stage. Then process carried out similarity these avoid oversegmentation. proposed homogram-based (HOB) compared with histogram-based (HIB). experimental results demonstrate that HOB can find more effectively than HIB does, solve problem discriminating shading in images some extent.