Video Compression for Screen Recorded Sequences Following Eye Movements.

Area of Science:

• Computer Science
• Multimedia Engineering
• Human-Computer Interaction

Background:

• Internet video traffic has surged due to smartphones and tablets.
• High Efficiency Video Coding (HEVC) was developed to reduce bit rates by 50%.
• Increasing demands for higher resolutions necessitate further bit rate reduction.

Purpose of the Study:

• To present a new algorithm for bit rate reduction in screen recorded videos.
• To leverage perceptual video coding and eye-tracking data for enhanced compression.
• To minimize bit rates while maintaining perceived video quality.

Main Methods:

• An eye-tracking system was employed during video recording to identify viewer fixation points.
• A variable quantization parameter (QP) was applied, with lower QP near fixation points.
• The algorithm focuses on encoding areas around fixation points with higher fidelity.

Main Results:

• The proposed algorithm achieved up to 31.3% bit rate savings compared to standard HEVC.
• No significant impact on the perceived quality of the video was observed.
• Demonstrated the effectiveness of eye-tracking in perceptual video compression.

Conclusions:

• Perceptual video coding, guided by eye-tracking, offers a promising approach for efficient video compression.
• The developed algorithm effectively reduces bit rates for screen recorded content.
• This method provides a viable solution for managing the growing demands of video traffic.