Single-click interactive cell nucleus image segmentation: Boosting accuracy with enhanced exclusion map and refinement segmentation

Nuclei in cell images are characterized by small size, high quantity, and high density, often with overlapping phenomena. This makes interactive methods such as contour-based approaches unsuitable for interactive cell nucleus image segmentation tasks. To improve the segmentation accuracy of nuclei, this study proposes a single-click interactive cell nucleus image segmentation method based on enhanced exclusion map and refinement segmentation. The proposed method employs a simple single-click interaction mechanism to alleviate the user's interaction burden. It introduces an enhanced exclusion map module and a Refinement Segmentation Module based on focused interaction to improve the quality of interactive cell nucleus image segmentation. The code is available at https://github.com/AICFStar/IClick.