Abstract

Cap4Bridge bridges the text-video modality gap by using generated captions as cross-modal context, enriched through stochastic augmentation during training. The method improves robustness and generalization in text-video retrieval without requiring additional human supervision.