DualGate-Net: A Prior-Gated Dual-Encoder Framework for Histopathology Cell Detection

2026-06-05

AI Summary

This work addresses the challenge that cell classification in histopathological images is highly sensitive to variations in the tissue microenvironment, and existing methods often introduce noise due to static feature fusion strategies. To overcome this limitation, the authors propose DualGate-Net, which integrates a ConvNeXtV2-based local encoder with a SegFormer-based global encoder and introduces a learnable spatially adaptive prior gating mechanism to dynamically fuse contextual tissue information. Additionally, a foreground reconstruction branch is incorporated to preserve high-frequency cellular structures, while a cellness-guided module enhances localization robustness. Evaluated on the OCELOT benchmark, the proposed method achieves macro F1-scores of 0.7722 and 0.7345 on the validation and test sets, respectively, significantly outperforming current state-of-the-art approaches.

Abstract

Cell detection in histopathology images strongly depends on surrounding tissue context, where visually similar cells may belong to different classes under different microenvironments. Recent tissue-aware methods incorporate contextual priors, but often rely on static fusion strategies that may propagate noisy information. In this work, we propose DualGate-Net, a prior-aware dual-encoder framework that combines a ConvNeXtV2-based local encoder and a SegFormer-based global encoder through a learnable prior-gated fusion mechanism. The proposed module adaptively regulates the influence of tissue priors across spatial locations, while an auxiliary foreground reconstruction branch preserves high-frequency cellular structures during training. In addition, auxiliary cellness-guided cues are incorporated to further improve localization robustness. Experiments on the OCELOT benchmark demonstrate consistent improvements, achieving macro F1-scores of 0.7722 on the validation set and 0.7345 on the test set, highlighting the effectiveness of adaptive prior integration for robust histopathology cell detection.

Problem

• histopathology cell detection
• tissue context
• contextual priors
• noisy information
• cell classification

Innovation

• prior-gated fusion
• dual-encoder architecture
• adaptive context integration
• foreground reconstruction
• cellness-guided cues