HVD: Human Vision-Driven Video Representation Learning for Text-Video Retrieval
This addresses the challenge of improving retrieval accuracy in text-video systems, though it appears incremental by building on CLIP-based methods.
The paper tackles the problem of 'blind' feature interaction in text-video retrieval, where models struggle to discern key visual information from background noise, by proposing the Human Vision-Driven (HVD) model with a coarse-to-fine alignment mechanism, achieving state-of-the-art performance on five benchmarks.
The success of CLIP has driven substantial progress in text-video retrieval. However, current methods often suffer from "blind" feature interaction, where the model struggles to discern key visual information from background noise due to the sparsity of textual queries. To bridge this gap, we draw inspiration from human cognitive behavior and propose the Human Vision-Driven (HVD) model. Our framework establishes a coarse-to-fine alignment mechanism comprising two key components: the Frame Features Selection Module (FFSM) and the Patch Features Compression Module (PFCM). FFSM mimics the human macro-perception ability by selecting key frames to eliminate temporal redundancy. Subsequently, PFCM simulates micro-perception by aggregating patch features into salient visual entities through an advanced attention mechanism, enabling precise entity-level matching. Extensive experiments on five benchmarks demonstrate that HVD not only captures human-like visual focus but also achieves state-of-the-art performance.