Gaze correction using feature-based view morphing and performance evaluation. Issue 1 (January 2017)

• Record Type: Journal Article
• Authors: Baek, Eu-Tteum; Ho, Yo-Sung
• Is Part Of: Signal, image and video processing. Volume 11:Issue 1(2017)
• Page Start: 187
• Page End: 194
• Publication Date: 2017-01
• Subjects: Eye gaze -- Face morphing -- Image synthesis -- Video conferencing; Signal processing -- Digital techniques -- Periodicals; Image processing -- Digital techniques -- Periodicals; Digital video -- Periodicals; 621.3822
• DOI: 10.1007/s11760-016-0918-1
• Languages: English
• ISSNs: 1863-1703

Abstract

Abstract Lack of eye gaze in the video conferencing system severely hinders immersive and comfortable communication. In this paper, we propose an eye gaze correction system for immersive video conferencing. Our system consists of a full HD display and two cameras mounted on the top and bottom of the display. In order to correct the eye gaze, we warp two captured images from their original positions to a virtual position, respectively. After we implement view morphing between corresponding triangles, we obtain a gaze-corrected image by interpolating the two morphed images. We also present a subjective evaluation method to measure the performance of our approach. Experimental results show that our system is effective in generating natural gaze-corrected images.