Robust Mitosis Detection Using a Cascade Mask-RCNN Approach With Domain-Specific Residual Cycle-GAN Data Augmentation

TL;DR

This paper presents a robust mitosis detection method using a cascade Mask-RCNN with domain-specific residual Cycle-GAN data augmentation to improve generalization across different scanner domains.

Contribution

The authors introduce a novel combination of a cascade Mask-RCNN detector with Cycle-GAN based domain augmentation and mitosis-specific bounding boxes for enhanced detection accuracy.

Findings

Achieved an F1 score of 0.7578 on the MIDOG challenge test set.

Second place in the MIDOG mitosis detection leaderboard.

Demonstrated improved generalization across scanner domains using Cycle-GAN augmentation.

Abstract

For the MIDOG mitosis detection challenge, we created a cascade algorithm consisting of a Mask-RCNN detector, followed by a classification ensemble consisting of ResNet50 and DenseNet201 to refine detected mitotic candidates. The MIDOG training data consists of 200 frames originating from four scanners, three of which are annotated for mitotic instances with centroid annotations. Our main algorithmic choices are as follows: first, to enhance the generalizability of our detector and classification networks, we use a state-of-the-art residual Cycle-GAN to transform each scanner domain to every other scanner domain. During training, we then randomly load, for each image, one of the four domains. In this way, our networks can learn from the fourth non-annotated scanner domain even if we don't have annotations for it. Second, for training the detector network, rather than using centroid-based fixed-size bounding boxes, we create mitosis-specific bounding boxes. We do this by manually annotating a small selection of mitoses, training a Mask-RCNN on this small dataset, and applying it to the rest of the data to obtain full annotations. We trained the follow-up classification ensemble using only the challenge-provided positive and hard-negative examples. On the preliminary test set, the algorithm scores an F1 score of 0.7578, putting us as the second-place team on the leaderboard.