Automatic chromosome G-band segmentation from background images

Abstract

The problem of chromosome segmentation is due to poor image quality, particularly lack of contrast and hole appeared in chromosome images. These images must be first enhanced and holes are to be filled up in the subsequent step. This research presents the technique developed for human chromosome image segmentation, base on Histogram Equalization, Adaptive Threshold, Flood-Fill, Canny Edge Detection and Axis-Aligned Bounding Boxes. Under this approach, HAFCO technique can improve enhancement and fill hole of chromosome images. From our study, we found that HAFCO give promising results in segmentation. A success rate of 97.27% has been achieved.