Abstract
In this paper we investigate the feasibility of using a wavelet transform to compress individual frames in a video sequence controlled by points of gaze coordinates collected from different viewers by an eye-tracker. Unlike most eye-tracking controlled compression approaches, we make the compression "off-line" and not in real time. Our results indicate that the distribution of gaze-points is "compact" across viewers in the sense that they tend to group together within limited regions. It is also observed that gaze-points collected from the original video sequence to a high degree cohere with those collected from the compressed sequence. Thus, wavelet-based foveated frame-sequence compression from eye-tracking data is possible, and does not change the viewing behavior of test-subjects.
| Original language | English |
|---|---|
| Title of host publication | Picture Coding Symposium 2004 |
| Publisher | PCS - Tektronix, Inc |
| Pages | 267-272 |
| Publication status | Published - 2004 |
| Event | Picture Coding Symposium 2004 - San Francisco, CA, United States Duration: 2004 Dec 15 → 2004 Dec 17 |
Conference
| Conference | Picture Coding Symposium 2004 |
|---|---|
| Country/Territory | United States |
| City | San Francisco, CA |
| Period | 2004/12/15 → 2004/12/17 |
Subject classification (UKÄ)
- Electrical Engineering, Electronic Engineering, Information Engineering
Free keywords
- Eye-tracking measurements
- Video sequences
- Gaze-point distribution
- Off-line foveation