In this work we present an approach to capture the total semantics in multimedia-multimodal web pages. Our research improves upon the state-ofthe-art with two key features: (1) cap...
We present methods for improving text search retrieval of visual multimedia content by applying a set of visual models of semantic concepts from a lexicon of concepts deemed relev...
Alexander Haubold, Apostol Natsev, Milind R. Napha...
This paper presents an attempt at building a large scale distributed composite language model that simultaneously accounts for local word lexical information, mid-range sentence s...
Many documents on the Web are formated in a weakly structured format. Because of their weak semantic and because of the heterogeneity of their formats, the information conveyed by...
Automatic image annotation is a promising solution to enable semantic image retrieval via keywords. In this paper, we propose a multi-level approach to annotate the semantics of n...