Stain-Robust Mitotic Figure Detection for the Mitosis Domain Generalization Challenge
This work addresses the challenge of domain generalization for mitotic figure detection, which is important for clinical applications but is incremental in nature.
The paper tackled the problem of detecting mitotic figures across different scanners/sites to assist clinicians with tumor grading, achieving an F1-score of 0.786 on training images and 0.765 on a preliminary test set.
The detection of mitotic figures from different scanners/sites remains an important topic of research, owing to its potential in assisting clinicians with tumour grading. The MItosis DOmain Generalization (MIDOG) challenge aims to test the robustness of detection models on unseen data from multiple scanners for this task. We present a short summary of the approach employed by the TIA Centre team to address this challenge. Our approach is based on a hybrid detection model, where mitotic candidates are segmented on stain normalised images, before being refined by a deep learning classifier. Cross-validation on the training images achieved the F1-score of 0.786 and 0.765 on the preliminary test set, demonstrating the generalizability of our model to unseen data from new scanners.