About This Topic
Develop a desktop app that renders a written text as follows: (1) TTS by an existing component, (2) Illustration based on the semantics in the text and outstanding terms. For example, if the text is a poem about a tree, one should be shown whenever the word "tree" occurs. In a fairy tale, "witch", "dwarf", etc. should be illustrated properly. Based on the simple association of terms with images (ideally, dynamically selected from a result set of a Google image query), more sophisticated linguistic concepts should be recognized and illustrated. E.g.: "the policeman beats the robber" Etc. Requires good knowledge in (visual) information retrieval.