RC-Net: Regression Correction for End-to-End Chromosome Instance Segmentation

Abstract

Precise segmentation of chromosome in the real image achieved by microscope is significant for karyotype analysis. The segmentation of image is usually achieved by pixel-level classification task, which consider different instances as different classes. Many instance segmentation methods predict the Intersection over Union (IoU) through the head branch to correct the classification confidence. Their effectiveness is based on the correlation between branch tasks. However, none of these methods consider the correlation between input and output in branch tasks. Herein, we propose a chromosome instance segmentation network based on regression correction, which adopts more relevant head branch between tasks to predict more relevant confidences with the positioning accuracy and segmentation accuracy to improve the classification confidence for segmentation. Furthermore, a non-maximum suppression algorithm is further designed to screen the target segmentation mask with the IoU of the overlapping instance. What’s more, given the fact that the original IoU loss function is not sensitive to the wrong segmentation, K-IoU loss function is defined to strengthen the penalty of the wrong segmentation and effectively improve the segmentation performance. Finally, an ablation experiment is designed to evaluate the effectiveness of the chromosome instance segmentation network based on regression correction, which shows that our proposed method can effectively enhance the performance in automatic chromosome segmentation tasks.

Keywords: Karyotype, Instance segmentation, Mask-based NMS, Correction