VTAgent: Agentic Keyframe Anchoring for Evidence-Aware Video TextVQA
For researchers in video understanding and VQA, this work provides a simple yet effective method to improve performance by focusing on evidence localization, though the approach is incremental.
The paper identifies that the bottleneck in Video TextVQA is keyframe localization, not reasoning, and proposes a question-guided agent framework that anchors keyframes before answering, achieving +12.12 accuracy and +11.15 ANLS improvement, setting new SOTA.
Video text-based visual question answering (Video TextVQA) aims to answer questions by reasoning over visual textual content appearing in videos. Despite the strong multimodal video understanding capabilities of recent Video-LLMs, their performance on existing Video TextVQA benchmarks remains limited. To better understand this gap, we conduct an upper-bound analysis through frame-wise question answering, counting a sample as correct if any frame yields the right answer, which significantly outperforms direct video-based inference and reveals a substantial performance gap. The results suggest that the primary bottleneck lies in the localization of key question-relevant evidence, rather than in reasoning capacity itself. Building on this insight, we propose a question-guided agent framework that explicitly anchors the relevant keyframes before answering. The approach operates effectively in a training-free setting and consistently surpasses direct video inference. With additional supervised fine-tuning (SFT) and reinforcement learning (RL), it achieves an average improvement of +12.12 in accuracy and +11.15 in ANLS across benchmarks, establishing new state-of-the-art results. Our study underscores the critical role of explicit keyframe anchoring for advancing Video TextVQA. The code will be publicly released.