Adversarial Reconstruction Feedback For Robust Fine-grained Generalization

Abstract

Existing fine-grained image retrieval (FGIR) methods predominantly rely on supervision from predefined categories to learn discriminative representations for retrieving fine-grained objects. However, they inadvertently introduce category-specific semantics into the retrieval representation, creating semantic dependencies on predefined classes that critically hinder generalization to unseen categories. To tackle this, we propose AdvRF, a novel adversarial reconstruction feedback framework aimed at learning category-agnostic discrepancy representations. Specifically, AdvRF reformulates FGIR as a visual discrepancy reconstruction task via synergizing category-aware discrepancy localization from retrieval models with category-agnostic feature learning from reconstruction models. The reconstruction model exposes residual discrepancies overlooked by the retrieval model, forcing it to improve localization accuracy, while the refined signals from the retrieval model guide the reconstruction mo

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keywang2025adversarial

Related papers

• DVF: Advancing Robust And Accurate Fine-grained Image Retrieval With Retrieval Guidelines (2024)9.03
• Fine-grained Image Retrieval Via Dual-vision Adaptation (2025)0.00
• Adaptive Fine-grained Sketch-based Image Retrieval (2022)9.76
• Language-driven Fine-grained Retrieval (2025)0.00
• Deep Reinforced Attention Regression For Partial Sketch Based Image Retrieval (2021)5.24
• A Study On FGSM Adversarial Training For Neural Retrieval (2023)7.81
• Coarse-to-fine: Learning Compact Discriminative Representation For Single-stage Image Retrieval (2023)9.35
• Refir: Grounding Large Restoration Models With Retrieval Augmentation (2024)0.00