Cascade RCNN for MIDOG Challenge

TL;DR

This paper introduces a multi-stage Cascade RCNN for mitosis detection in breast cancer images, improving false positive rejection and achieving an F1-score of 0.7492 on a test set.

Contribution

The work presents a novel multi-stage Cascade RCNN approach specifically designed for mitosis detection, enhancing selectivity against false positives.

Findings

Achieved an F1-score of 0.7492 on the test set.

Demonstrated improved false positive rejection with the cascade approach.

Applicable to breast cancer prognosis imaging.

Abstract

Mitotic counts are one of the key indicators of breast cancer prognosis. However, accurate mitotic cell counting is still a difficult problem and is labourious. Automated methods have been proposed for this task, but are usually dependent on the training images and show poor performance on unseen domains. In this work, we present a multi-stage mitosis detection method based on a Cascade RCNN developed to be sequentially more selective against false positives. On the preliminary test set, the algorithm scores an F1-score of 0.7492.