Contrastive Learning of Semantic and Visual Representations for Text Tracking
This work addresses the problem of robustly tracking text in videos for applications like video analysis, offering a domain-specific incremental improvement.
The paper tackles video text tracking by integrating semantic and visual representations using contrastive learning, resulting in an end-to-end tracker (SVRep) that achieves a state-of-the-art IDF1 of 65.9% with 16.7 FPS on the ICDAR2015 dataset, an 8.6% improvement over previous methods.
Semantic representation is of great benefit to the video text tracking(VTT) task that requires simultaneously classifying, detecting, and tracking texts in the video. Most existing approaches tackle this task by appearance similarity in continuous frames, while ignoring the abundant semantic features. In this paper, we explore to robustly track video text with contrastive learning of semantic and visual representations. Correspondingly, we present an end-to-end video text tracker with Semantic and Visual Representations(SVRep), which detects and tracks texts by exploiting the visual and semantic relationships between different texts in a video sequence. Besides, with a light-weight architecture, SVRep achieves state-of-the-art performance while maintaining competitive inference speed. Specifically, with a backbone of ResNet-18, SVRep achieves an ${\rm ID_{F1}}$ of $\textbf{65.9\%}$, running at $\textbf{16.7}$ FPS, on the ICDAR2015(video) dataset with $\textbf{8.6\%}$ improvement than the previous state-of-the-art methods.