Instance Segmentation of Multiple Myeloma Cells via Hybrid Task Cascade

Abstract

Multiple Myeloma (MM) is a blood cancer that develops when plasma cells expand abnormally in the bone marrow. Early detection of MM is beneficial for accurate treatment in time and draws increasing recognition. There are several endeavors to construct computer-assisted automatic diagnostic tools for myeloma cell detection. Of late, deep learning based methods have been expected to apply for the detection and segmentation of cells of interest from microscopic images. In this paper, an ensemble framework is designed for the detection and segmentation of myeloma cells. Extensive experimental results on the SegPC-2021 Challenge dataset demonstrate that the proposed method achieves promising performance.

Tasks

Reproductions