TranX-Adapter: Bridging Artifacts and Semantics within MLLMs for Robust AI-generated Image Detection

Rapid advances in AI-generated image (AIGI) technology enable highly realistic synthesis, threatening public information integrity and security. Recent studies have demonstrated that incorporating texture-level artifact features alongside semantic features into multimodal large language models (MLLMs) can enhance their AIGI detection capability. However, our preliminary analyses reveal that artifact features exhibit high intra-feature similarity, leading to an almost uniform attention map after the softmax operation. This phenomenon causes attention dilution, thereby hindering effective fusion between semantic and artifact features. To overcome this limitation, we propose a lightweight fusion adapter, TranX-Adapter, which integrates a Task-aware Optimal-Transport Fusion that leverages the Jensen-Shannon divergence between artifact and semantic prediction probabilities as a cost matrix to transfer artifact information into semantic features, and an X-Fusion that employs cross-attention to transfer semantic information into artifact features. Experiments on standard AIGI detection benchmarks upon several advanced MLLMs, show that our TranX-Adapter brings consistent and significant improvements (up to +6% accuracy).

Key Contributions

• Identifies 'attention dilution' phenomenon that degrades artifact-semantic feature fusion in MLLMs due to high intra-feature similarity of artifact representations
• Proposes TOP-Fusion (Task-Aware Optimal-Transport Fusion) using Jensen-Shannon divergence as a cost matrix to transfer artifact information into semantic features without attention collapse
• Proposes X-Fusion using cross-attention for the reverse direction (semantic→artifact), enabling bidirectional feature interaction with up to +6% accuracy gain on AIGI detection benchmarks

🛡️ Threat Analysis

Details

Similar Papers