The Department of Computer Science at the University of Cyprus cordially invites you to the PhD Defense entitled:
Real-time Adaptation to Time-Varying Constraints for Reliable mHealth Video Communications
Speaker: Mr. Zinonas C. Antoniou
The significant number of mobile health (mHealth) systems developed over the past decade reflect the broad applicability spectrum of such systems and services in standard clinical practice. In terms of medical video communications, application scenarios range –but are not limited- from remote diagnosis and care, to emergency response, 2nd opinion provision, and home monitoring. Integration of medical video communication systems in the healthcare provision pathway can significantly enhance the quality of care while reducing hospitalization times and associated healthcare costs. The challenge lies in delivering sufficiently high video resolutions and frame rates with the low-delay and low packet loss rates requirements that will accommodate an equivalent level of clinical experience to that of in-hospital examinations. The latter requires real-time control of the video streaming process so as to facilitate the adequate levels of clinical video quality required to support reliable diagnosis. The motivation of this PhD thesis is to provide a scalable, video modality, encoder, and wireless network agnostic framework, that will support real-time adaptation to time-varying wireless networks’ state while guaranteeing diagnostically lossless video communications and conforming to end-user device constraints for real-time performance. More specifically, the approach is based on multi-objective optimization, that jointly maximizes the encoded video’s quality and encoding rate while minimizing bitrate demands. For this purpose, a dense encoding space is constructed and linear regression is used to estimate forward prediction models for quality, bitrate, and computational complexity. The prediction models are then used by the proposed adaptive control framework that can fine-tune video encoding based on real-time constraints. The proposed framework is validated using a leave-one-out algorithm applied to ten ultrasound videos of the common carotid artery. The prediction models can estimate Structural SIMilarity (SSIM) quality with a median accuracy error of less than 1%, bitrate demands with deviation error of 10% or less, and encoding frame rate within a 6% margin. Real-time adaptation at a Group of Pictures (GOP) level is demonstrated using the High Efficiency Video Coding (HEVC) standard. The effectiveness of the proposed framework compared to static, non-adaptive approaches is demonstrated for different modes of operation, achieving significant quality gains, bitrate demands reductions, and performance improvements, in real-life scenarios imposing time-varying constraints.
Zinonas C. Antoniou is a Ph.D. candidate at the Department of Computer Science under the supervision of Professor Constantinos S. Pattichis and co-supervision of Dr. Andreas S. Panayides. He received his 5-year Engineering Diploma (Dipl.-Ing.) from the School of Electrical and Computer Engineering, of the National Technical University of Athens. His major research interests include video processing and communications, mHealth applications, and mobile telecommunication networks. He worked as Teaching Assistant in various undergraduate courses at the Department of Computer Science and Business and at Department of Business and Public Administration of the University of Cyprus. He also worked as a Research Assistant at the Department of Computer Science of University of Cyprus for four EU funded projects, namely GRANATUM, FI-STAR, e-ENERCA and CEF. He is a member of the eHealth Laboratory and he is currently a Special Scientist at the Department of Computer Science of University of Cyprus.
|Other Presentations Web: https://www.cs.ucy.ac.cy/colloquium/presentations.php|
|Colloquia Web: https://www.cs.ucy.ac.cy/colloquium/|