COIN: Confidence Score-Guided Distillation for Annotation-Free Cell Segmentation

Sanghyun Jo * , Seo Jin Lee * , Seungwoo Lee , Seohyung Hong , Hyungseok Seo † , Kyungsu Kim †
OGQ
Seoul National University
ICCV 2025
* Equal contribution
† Corresponding author
Paper Code arXiv
MY ALT TEXT
In this paper, we present COIN, a three-step approach that overcomes the absence of error-free instances, inherent in existing UCIS methods, through instance-level confidence scoring approach combined with recursive self-distillation. Notably, our method achieves substantial performance improvement in instance segmentation by more than twofold in SSA and at least +18%p in PSM on MoNuSeg. Our extensive experiments demonstrate that while COIN operates without relying on any image-related annotations, it consistently outperforms supervised models.

Abstract

Cell instance segmentation (CIS) is crucial for identifying individual cell morphologies in histopathological images, providing valuable insights for biological and medical research. While unsupervised CIS (UCIS) models aim to reduce the heavy reliance on labor-intensive image annotations, they fail to accurately capture cell boundaries, causing missed detections and poor performance. Recognizing the absence of error-free instances as a key limitation, we present COIN (COnfidence score-guided INstance distillation), a novel annotation-free framework with three key steps: (1) Increasing the sensitivity for the presence of error-free instances via unsupervised semantic segmentation with optimal transport, leveraging its ability to discriminate spatially minor instances, (2) Instance-level confidence scoring to measure the consistency between model prediction and refined mask and identify highly confident instances, offering an alternative to ground truth annotations, and (3) Progressive expansion of confidence with recursive self-distillation. Extensive experiments across six datasets show COIN outperforming existing UCIS methods, even surpassing semi- and weakly-supervised approaches across all metrics on the MoNuSeg and TNBC datasets.

Problem

MY ALT TEXT

Recently developed unsupervised CIS (UCIS) methods focus on extracting the most discriminative parts with geometric augmentation (e.g., rotation), which fails to address edge-discriminative features. Therefore, as shown in figure, we observe that none of recent UCIS methods could predict even a single instance with 100% accuracy, a phenomenon we define as the absence of error-free instances.

Method

MY ALT TEXT

COIN is divided into three steps:

  1. Pixel-level cell propagation: allows the detection of all cells with high sensitivity and ensures the presence of an error-free instance.
  2. Instance-level confidence scoring: we assess the confidence of each instance by applying an unsupervised scoring approach to extract error-free instances.
  3. Confidence score-guided recursive self-distillation: occurs with the guidance from (2), allowing progressive increment in the number of confident cells.

Experiment

BibTeX

@InProceedings{jo2025coin,
      title={COIN: Confidence Score-Guided Distillation for Annotation-Free Cell Segmentation}, 
      author={Sanghyun Jo and Seo Jin Lee and Seungwoo Lee and Seohyung Hong and Hyungseok Seo and Kyungsu Kim},
      booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
      year={2025}
}