Abstract:
This research focuses on the development of an automated number determination method for amniotic fluid chromosome images (Automated Number Determination Development for Amniotic Fluid Bag Chromosomes Image: AND-AFBCI).
The aim is to enhance the accuracy and efficiency of automated chromosome counting by addressing the issue of erroneous crosspoint detection, which arises due to chromosome image ambiguities caused by overlapping and bending chromosomes.
These issues often result in incorrect chromosome counts. Therefore, this research
seeks to develop a novel crosspoint identification method using the Crosspoint
Identification in Chromosomal Structures: A Directional Binary Approach (CICS-DBA) and a Hybrid Approach for Chromosome Segmentation in Amniotic Fluid Cells Using OTSU, Flood Fill, and Morphological Techniques (HCS-AFC). The methods are built on Direct Binary-Scale images and the Hit-or-Miss Transformation technique.
The experimental results show that the AND-AFBCI method improves the accuracy of chromosome counting in overlapping amniotic fluid chromosome images, achieving an accuracy rate of 81.48%.