Targeted False Positive Synthesis via Detector-guided Adversarial Diffusion Attacker for Robust Polyp Detection

2025-06-22

AI Summary

Polyp detection models are highly susceptible to false positives, yet existing data augmentation techniques primarily emphasize diversity enhancement while overlooking the constructive utility of synthesizing false-positive samples. To address this, we propose DADA (Detection-guided Adversarial Diffusion Attack), the first adversarial diffusion framework explicitly designed for improving detector robustness. DADA innovatively introduces adversarial diffusion into lesion detection by jointly leveraging region-wise noise matching and detector gradient-guided perturbations to generate highly confounding false positives in a targeted manner. The method integrates negative-sample-centered diffusion modeling with generative adversarial mechanisms, substantially enhancing augmentation efficacy. Evaluated on both public and internal colorectal cancer screening datasets, DADA-synthesized data improves detector F1 scores by ≥2.6% and ≥2.7%, respectively, demonstrating its effectiveness in suppressing false positives and strengthening generalization capability.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Search and Optimization: Adversarial Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Polyp detection is crucial for colorectal cancer screening, yet existing models are limited by the scale and diversity of available data. While generative models show promise for data augmentation, current methods mainly focus on enhancing polyp diversity, often overlooking the critical issue of false positives. In this paper, we address this gap by proposing an adversarial diffusion framework to synthesize high-value false positives. The extensive variability of negative backgrounds presents a significant challenge in false positive synthesis. To overcome this, we introduce two key innovations: First, we design a regional noise matching strategy to construct a negative synthesis space using polyp detection datasets. This strategy trains a negative-centric diffusion model by masking polyp regions, ensuring the model focuses exclusively on learning diverse background patterns. Second, we introduce the Detector-guided Adversarial Diffusion Attacker (DADA) module, which perturbs the negative synthesis process to disrupt a pre-trained detector's decision, guiding the negative-centric diffusion model to generate high-value, detector-confusing false positives instead of low-value, ordinary backgrounds. Our approach is the first to apply adversarial diffusion to lesion detection, establishing a new paradigm for targeted false positive synthesis and paving the way for more reliable clinical applications in colorectal cancer screening. Extensive results on public and in-house datasets verify the superiority of our method over the current state-of-the-arts, with our synthesized data improving the detectors by at least 2.6% and 2.7% in F1-score, respectively, over the baselines. Codes are at https://github.com/Huster-Hq/DADA.

Problem

• Enhancing polyp detection by synthesizing high-value false positives
• Addressing limited data diversity in colorectal cancer screening models
• Improving detector robustness via adversarial diffusion framework

Innovation

• Regional noise matching for negative synthesis space
• Detector-guided Adversarial Diffusion Attacker (DADA)
• Adversarial diffusion for lesion detection