SSIM-Based Distortion Estimation for Optimized Video Transmission over Inherently Noisy Channels

Abstract

The authors present two methods for examining video quality using the Structural Similarity (SSIM) index: Iterative Distortion Estimate (IDE) and Cumulative Distortion using SSIM (CDSSIM). In the first method, three types of slices are iteratively reconstructed frame-by-frame for three different combinations of packet loss and the resulting distortions are combined using their probabilities to give the total expected distortion. In the second method, a cumulative measure of the overall distortion is computed by summing the inter-frame propagation impact to all frames affected by a slice loss. Furthermore, the authors develop a No-Reference (NR) sparse regression framework for predicting the CDSSIM metric to circumvent the real-time computational complexity in streaming video applications. The two methods are evaluated in resource allocation and packet prioritization schemes and experimental results show improved performance and better end-user quality. The accuracy of the predicted CDSSIM values is studied using standard performance measures and a Quartile-Based Prioritization (QBP) scheme.

A novel texture-based asymmetric visibility threshold model for stereoscopic video coding

“…Not only the classical objective quality evaluation metric PSNR but also structural similarity index measurement (SSIM) is used to evaluate video quality in this paper. Compared with PSNR, SSIM is found to be a better indicator of perceived image quality than mean-squared error which is the theoretical basis of PSNR [28,29]. PSNR and SSIM value of six reconstructed video sequences with different QP are shown in Table 13.…”