Cineast: a Content-based Video Retrieval Engine (Master Thesis, Finished)


Luca Rossetto


The recent increase in availability of cheap video recording equipment and processing power has led to a massive increase in the amount of videos stored. Over the last decade or two, research have come up with efficient ways for text retrieval, however, there are still a lot of open questions in the retrieval of multimedia data such as video. The Cineast project will develop and evaluate a content-based retrieval system for videos. This project consists of three major parts: First, the compilation of a general purpose short film collection for experimentation and evaluation purposes. Second, the implementation of a content-based video retrieval engine based on ADAM, that is capable of using various features for similarity retrieval. Third, the evaluation of the retrieval quality of the single features implemented.

Start / End Dates

2013/11/13 - 2014/05/12


Research Topics