Visual question answering: Which investigated applications?

Visual Question Answering (VQA) is an extremely stimulating and challenging research area where Computer Vision (CV) Natural Language Processig (NLP) have recently met. In image captioning video summarization, the semantic information completely contained in still images or dynamics, it has only to be mined expressed a human-consistent way. Differently from this, VQA same media must compared with semantics implied by question natural language, doubling artificial intelligence-related effort. Some recent surveys about approaches focused on methods underlying either image-related processing verbal-related one, way consistently fuse conveyed information. Possible applications are suggested, and, fact, most cited works rely general-purpose datasets that used assess building blocks of system. This paper rather considers proposals focus real-world applications, possibly using as benchmarks suitable data bound application domain. The also reports some challenges research.