Position-based anchor optimization for point supervised dense nuclei detection
Abstract
Nuclei detection is one of the most fundamental and challenging problems in histopathological image analysis, which can localize nuclei to provide effective computer-aided cancer diagnosis, treatment decision, and prognosis. The fully-supervised nuclei detector requires a large number of nuclei annotations on high-resolution digital images, which is time-consuming and needs human annotators with professional knowledge.
In recent years, weakly-supervised learning has attracted significant attention in reducing the labeling burden. However, detecting dense nuclei of complex crowded distribution and diverse appearances remains a challenge.
Methodology
To solve this problem, we propose a novel point-supervised dense nuclei detection framework that introduces position-based anchor optimization to complete morphology-based pseudo-label supervision.
1. Cellular-level Pseudo Labels (CPL): We first generate cellular-level pseudo labels for the detection head via a morphology-based mechanism, which can help to build a baseline point-supervised detection network.
2. Position-based Anchor-quality Estimation (PAE): Considering the crowded distribution of the dense nuclei, we propose PAE mechanism, which utilizes the positional deviation between an anchor and its corresponding point label to suppress low-quality detections far from each nucleus.
3. Adaptive Anchor Selector (AAS): To better handle the diverse appearances of nuclei, an AAS operation is proposed to automatically select positive and negative anchors according to morphological and positional statistical characteristics of nuclei.
Experimental Results
We conduct comprehensive experiments on two widely used benchmarks, MO and Lizard, using ResNet50 and PVTv2 as backbones. The results demonstrate that the proposed approach has superior capacity compared with other state-of-the-art methods.
In particularly, in dense nuclei scenarios, our method can achieve 95.1% performance of the fully-supervised approach, significantly reducing the annotation burden while maintaining high detection accuracy.
The code is available at https://github.com/NucleiDet/DenseNucleiDet.