REVIVE: Regional Visual Representation Matters in Knowledge-Based Visual Question Answering
This work addresses a bottleneck in knowledge-based VQA for AI systems by improving visual representation, though it is incremental as it builds on existing methods with a focus on regional features.
The paper tackles the problem of underutilized regional visual information in knowledge-based visual question answering (VQA) by proposing REVIVE, a method that integrates object region details in both knowledge retrieval and answering stages, achieving a state-of-the-art accuracy of 58.0% on the OK-VQA dataset, a 3.6% improvement over previous methods.
This paper revisits visual representation in knowledge-based visual question answering (VQA) and demonstrates that using regional information in a better way can significantly improve the performance. While visual representation is extensively studied in traditional VQA, it is under-explored in knowledge-based VQA even though these two tasks share the common spirit, i.e., rely on visual input to answer the question. Specifically, we observe that in most state-of-the-art knowledge-based VQA methods: 1) visual features are extracted either from the whole image or in a sliding window manner for retrieving knowledge, and the important relationship within/among object regions is neglected; 2) visual features are not well utilized in the final answering model, which is counter-intuitive to some extent. Based on these observations, we propose a new knowledge-based VQA method REVIVE, which tries to utilize the explicit information of object regions not only in the knowledge retrieval stage but also in the answering model. The key motivation is that object regions and inherent relationship are important for knowledge-based VQA. We perform extensive experiments on the standard OK-VQA dataset and achieve new state-of-the-art performance, i.e., 58.0% accuracy, surpassing previous state-of-the-art method by a large margin (+3.6%). We also conduct detailed analysis and show the necessity of regional information in different framework components for knowledge-based VQA. Code is publicly available at https://github.com/yzleroy/REVIVE.