Adaptive Image Enhancement for Automatic Complicated G-band Chromosome Number Determination

The light microscopic image of the chromosome is one of the sources to diagnose genetic disorders. Chromosome counting is the first step for diagnosing genetic abnormalities. However, in the case of complicated chromosome pattern in images, it still needs improvement due to its complication and poor image quality. Moreover, every chromosome images are different in contrast and brightness. Therefore, in order to achieve a better performance in chromosome counting, the complicated chromosome images need to be specially enhanced. This paper proposes a technique called Adaptive Complicated Chromosome Image Enhancement (ACCIE) using the determining threshold value method, and the dynamically determining intensity adjustment method. The proposed method helps to generate a proper chromosome skeletonization and yields 86.81% for complicated chromosome number determination.