Video Objects Description in Hindi Text Language
Vandana D. Edke, Ramesh Mahadev Kagalkar · International Journal of Managing Public Sector Information and Communication Technologies · 2016
Video activity recognition has grown to be a dynamic location of analysis in latest years.A widespread information-driven approach is denoted in this paper that produces descriptions of video content into textual content description inside the Hindi language.This method combines the final results of modern item with "real-international" records to pick the in all subject-verb-object triplet for depicting a video.The usage of this triplet desire technique, a video is tagged via the trainer, mainly, Subject, Verb, and object (SVO) and then this data is mined to improve the result of checking out video clarification by using pastime as well as item identity.Contrasting preceding approaches, this method can annotate arbitrary videos deprived of wanting the large series and annotation of a similar schooling video corpus.The proposed work affords initial and primary text description within the Hindi language that is producing easy words and sentence formation.But the fundamental challenging attempt on this work is to extract grammatically accurate and expressive text records in Hindi textual content regarding video content.