Unveil: Unified Visual-Textual Integration and Distillation for Multi-modal Document Retrieval
For practitioners needing robust document retrieval across diverse formats, this work offers a novel approach that combines visual and textual cues without relying on error-prone parsing, though the gains are incremental over existing methods.
Unveil proposes a visual-textual embedding framework for multi-modal document retrieval that integrates textual and visual features, then distills knowledge into a purely visual model. The method surpasses existing approaches in retrieval accuracy and efficiency, with knowledge distillation bridging the performance gap between visual-textual and visual-only methods.
Document retrieval in real-world scenarios faces significant challenges due to diverse document formats and modalities. Traditional text-based approaches rely on tailored parsing techniques that disregard layout information and are prone to errors, while recent parsing-free visual methods often struggle to capture fine-grained textual semantics in text-rich scenarios. To address these limitations, we propose \textbf{Unveil}, a novel visual-textual embedding framework that effectively integrates textual and visual features for robust document representation. Through knowledge distillation, we transfer the semantic understanding capabilities from the visual-textual embedding model to a purely visual model, enabling efficient parsing-free retrieval while preserving semantic fidelity. Experimental results demonstrate that our visual-textual embedding method surpasses existing approaches, while knowledge distillation successfully bridges the performance gap between visual-textual and visual-only methods, improving both retrieval accuracy and efficiency.