VACE2 Close out workshop

I am attending ‘Video Analysis Content Extraction’ (VACE) Phase 2 Closeout Workshop. VACE is an interesting and important program sponsored by different groups of the Department of Defense. This program recognizes the fact that enormous volumes of video are being generated in different contexts as varied as Broadcast News, video surveillance, meetings, reconnaissance, and unmanned aerial vehicles and remain unexplored due to unavailability of suitable tools. There is valuable information in all these videos that must be extracted to build systems that will allow analysts to find relevant information. Definitely this is a very important, worthwhile, and ambitious goal.

As the name suggests, the focus has been on techniques for visual analysis of data. Powerful techniques from many leading computer vision research groups, both from academia and industry in USA, presented their work funded under this project at this workshop. It was made clear in the introductory talk by Dennis Moellman, the program manager of the project, that one of the major accomplishments of the program is advancement of computer vision technology. Techniques for tracking, detecting visual events like abandoned objects, building 3-dimensional models from video, recognizing facial expressions, recognizing sign language, and recognizing multilingual text in video are making progress and are reported here.

Clearly many interesting techniques and their applications are presented at this workshop. It is impressive to see how computer vision techniques are improving primarily due to availability of powerful processing and storage environment.

This phase was still more oriented towards developing component techniques. The real challenge for building applications is analyzing the problems that need to be solved and then pull together appropriate component technologies and design systems that could use these techniques synergistically. So far, in the VACE program the emphasis has been on component technology – the applications are more for motivation or inspiration. The next phase of research is the final phase. Hopefully in this phase there will be emphasis on pulling these techniques together and develop real applications. Even contrived applications will provide lots of experience about the efficacy and limitations of the component technology.

Leave a Reply