DeNuC: Decoupling Nuclei Detection and Classification in Histopathology
This work provides a more efficient and accurate method for nuclei detection and classification in histopathology, which is significant for researchers and practitioners working with pathology image analysis.
This paper addresses the problem of suboptimal performance of Pathology Foundation Models (FMs) in Nuclei Detection and Classification (NDC) by identifying that joint optimization leads to representation degradation and computational burden. The proposed method, DeNuC, decouples these tasks, achieving F1 score improvements of 4.2% on BRCAM2C and 3.6% on PUMA datasets, while using 16% fewer trainable parameters.
Pathology Foundation Models (FMs) have shown strong performance across a wide range of pathology image representation and diagnostic tasks. However, FMs do not exhibit the expected performance advantage over traditional specialized models in Nuclei Detection and Classification (NDC). In this work, we reveal that jointly optimizing nuclei detection and classification leads to severe representation degradation in FMs. Moreover, we identify that the substantial intrinsic disparity in task difficulty between nuclei detection and nuclei classification renders joint NDC optimization unnecessarily computationally burdensome for the detection stage. To address these challenges, we propose DeNuC, a simple yet effective method designed to break through existing bottlenecks by Decoupling Nuclei detection and Classification. DeNuC employs a lightweight model for accurate nuclei localization, subsequently leveraging a pathology FM to encode input images and query nucleus-specific features based on the detected coordinates for classification. Extensive experiments on three widely used benchmarks demonstrate that DeNuC effectively unlocks the representational potential of FMs for NDC and significantly outperforms state-of-the-art methods. Notably, DeNuC improves F1 scores by 4.2% and 3.6% (or higher) on the BRCAM2C and PUMA datasets, respectively, while using only 16% (or fewer) trainable parameters compared to other methods. Code is available at https://github.com/ZijiangY1116/DeNuC.