M.E. Stachura, S.P. Rao, E.V. Khasanshina, A. Pearson-Shaver, and S.L. Robertson (USA)
Region-of-Interest, Mathematical-losslessness, Perceptual -losslessness, Diagnostic-losslessness, Best-effort, Encoder-state.
We present a methodology for delivering diagnostically acceptable video quality over limited and variable bitrate networks using a region-of-interest framework. The design of the system consists of three steps – quantification of video quality through tests on training videos with medical experts, incorporation of the above information into the video encoder through the notion of encoder states, and fine tuning of these states based on built-in quality criteria. We used videos of patients in respiratory distress to train and test our bit allocation algorithms. Since this is a work in progress, the results in this paper are preliminary, and give insight into the expected performance of our proposed elastic system.
Important Links:
Go Back
