Towards high quality region-of-interest medical video over wireless networks using lossless coding and motion compensated temporal filtering

Telehealth is growing rapidly and there is an increased need for high quality transmission of video over wireless networks. In this paper, we propose a method for achieving high quality within a diagnostically important region-of-interest (ROI) using mathematically lossless (M-lossless) video coding of the ROI and motion compensated temporal filtering (MCTF) of the video. The remaining region (non-ROI) is coded in a lossy manner. MCTF as a prefilter addresses the issue of noise in video. With the correct values of filter parameters, it results in very good denoising for noisy video and improved quality for even near noiseless video. The key motivation to use MCTF in our application is that in both cases it results in significant reduction of the required total bitrate. This is crucial on data rate constrained wireless networks. We also show how MCTF is particularly advantageous for the ROI. We study the advantages of MCTF in terms of (i) quality improvement, and (ii) bitrate reduction. We apply a mean MCTF to filter Additive White Gaussian Noise (AWGN), and propose a new median MCTF to filter impulse noise. The results show that the applied techniques result in substantial improvement in quality and significant total bitrate reduction