- No upcoming events available
Scalable Visual Analytics of Video Data
|
Applicant: |
Prof. Dr. Gunther Heidemann, Stuttgart Prof. Dr. Daniel Weiskopf, Stuttgart |
|
Project: |
Scalable Visual Analytics of Video Data |
|
Summary: |
A substantial amount of electronic data that is acquired worldwide on a regular basis is in the form of video data, for example, by closed circuit television (CCTV) cameras, video streams from websites, or from sports video. Often, video is accompanied by associated data, e.g., sound or position information from GPS sensors in mobile phones. Directly viewing video footage does not scale well with the amount of video because it permanently requires user attention even for, in most parts, uninteresting material. The main goal of this project is to support users in the interactive analysis of video data in combination with associated data in order to efficiently understand the structure of behavior and to identify "regular" and "irregular" behavior. One challenge is that "irregular" behavior cannot be completely defined beforehand and, therefore, fully automatic techniques cannot provide a complete analysis. Another challenge is the difficulty of interpreting complex time-dependent multivariate data that may not contain complete information (e.g., due to occlusion or insufficient recording quality). Our strategy is to combine partially automatic analysis with visualization and interaction, relying on the human users’ excellent capabilities of identifying structure and interpreting video with associated data. By integrating the applicants’ expertise in computer vision, intelligent systems, visualization, perception-oriented graphics, and interaction techniques, this project has the goal of a scalable visual analytics tool for multivariate video data that supports the complete analysis and reasoning process all the way to dissemination and communication of analysis results. |