This paper describes how the author used the Java Media Framework to implement some rudimentary dimensionality reduction techniques on videos, in order to make them more easily tractable for further processing. She targets two application scenarios: the classification and identification of videos according to their color and abridgment of video contents for navigation.

Unfortunately, the paper neither provides an introduction to programming with the Java Media Framework, nor an introduction to the statistical methods behind the formulas presented. Therefore, I am not sure of the paper’s target audience. The methods depicted are far from being state-of-the-art video retrieval and multimedia processing techniques. With more background and a fundamental explanation of the techniques and the programming, it might be suited for an undergraduate computer science student. As the paper is, however, I have a hard time recommending spending time on it.