Generalized Image Captioning for Multilingual Support

Image captioning is a problem of viewing images and describing in language. This an important that can be solved by understanding the image, combining two fields image processing natural language into one. The purpose research so far has been to create general explanatory captions learning data. However, various environments reality must considered for practical use, as well descriptions suit use. caption requires new data generate descriptive specific purposes, but it takes lot time effort learnable In this study, we propose method solve problem. Popular help visually impaired people understand their surroundings automatically recognizing text then voice issue applied many places such search, art therapy, sports commentary, real-time traffic information commentary. Through domain object dictionary proposed without need process adjusting each application. study change focus on rather than data, leading creation intensively explaining objects required domain. work, filter model induces generation from domains while maintaining performance existing models.