DPA-P2PNet: Deformable Proposal-aware P2PNet for Accurate Point-based Cell Detection

TL;DR

DPA-P2PNet is a novel point-based cell detection method that leverages multi-scale features and deformable proposals to improve accuracy and address scale and positional biases, with self-supervised pre-training enhancing performance.

Contribution

The paper introduces DPA-P2PNet, which incorporates multi-scale feature extraction and deformable proposals, along with a multi-field-of-view variant and self-supervised pre-training for improved cell detection.

Findings

Outperforms state-of-the-art methods on three benchmarks.

Effective multi-scale feature decoding improves localization accuracy.

Self-supervised pre-training enhances model performance.

Abstract

Point-based cell detection (PCD), which pursues high-performance cell sensing under low-cost data annotation, has garnered increased attention in computational pathology community. Unlike mainstream PCD methods that rely on intermediate density map representations, the Point-to-Point network (P2PNet) has recently emerged as an end-to-end solution for PCD, demonstrating impressive cell detection accuracy and efficiency. Nevertheless, P2PNet is limited to decoding from a single-level feature map due to the scale-agnostic property of point proposals, which is insufficient to leverage multi-scale information. Moreover, the spatial distribution of pre-set point proposals is biased from that of cells, leading to inaccurate cell localization. To lift these limitations, we present DPA-P2PNet in this work. The proposed method directly extracts multi-scale features for decoding according to the coordinates of point proposals on hierarchical feature maps. On this basis, we further devise deformable point proposals to mitigate the positional bias between proposals and potential cells to promote cell localization. Inspired by practical pathological diagnosis that usually combines high-level tissue structure and low-level cell morphology for accurate cell classification, we propose a multi-field-of-view (mFoV) variant of DPA-P2PNet to accommodate additional large FoV images with tissue information as model input. Finally, we execute the first self-supervised pre-training on immunohistochemistry histopathology image data and evaluate the suitability of four representative self-supervised methods on the PCD task. Experimental results on three benchmarks and a large-scale and real-world interval dataset demonstrate the superiority of our proposed models over the state-of-the-art counterparts. Codes and pre-trained weights will be available.