Visual Question Answering Based on Visual Contentand Query Enhancement

With the rapid development of computer vision and natural language processing technologies, visual question answering (VQA), as an intersection of these two fields, is gradually becoming a research hotspot. Existing VQA models have achieved significant progress in general scenarios, but in data-biased scenarios, methods for mitigating bias still suffer from poor transferability. Meanwhile, in recent years, multimodal information enhancement techniques have made remarkable progress in exploring and integrating cross-modal semantics, offering powerful cross-modal semantic representation capabilities that provide new solutions for addressing issues caused by data bias. Therefore, this paper takes multimodal information enhancement as the entry point and designs a VQA method based on visual content and query enhancement to meet the demands of data-biased scenarios, aiming to improve the model’s reasoning capabilities and overall performance in such contexts.