Extraction of karyocytes and their components from microscopic bone marrow images based on regional color features

摘要

Extracting karyocytes and their components from microscopic bone marrow images is prerequisite for computer-aided early diagnosis of hemopathy. Most existing methods assume all pixels belonging to a karyon region or a cytoplasm region have similar colors. Practically, the color of neither a karyon nor a cytoplasm in a microscopic image is homogeneous in the pixel level. Therefore, the regional color features of a region are emphasized in this paper. Based on this observation, we propose a novel method to karyocyte extraction that first identifies a karyon by 4-connected block growing from a karyon feature block, then identifies feature blocks of its cytoplasm based on the extracted karyon, and finally extracts all cytoplasm regions by growing the cytoplasm feature blocks. Combining the karyon region and the corresponding cytoplasm regions can attain a complete karyocyte. Experimental results show that the proposed method is effective and robust.