Kindai University
Multimedia Understanding Group

Large-scale video retrieval

We are developing a video retrieval system that can efficiently and accurately find videos relevant to a query from more than one million videos (the size is more than 1TB). In particular, we are exploring fully-automatic approaches that automate the analysis of contents in each video and the decision of whether it is relevant to a query.

The poster below is a high-resolution image. We recommend to display it in a separate window or tab, or download it.