Real-time 3D QoE evaluation of novel 3D media

Recent wireless networks enable the transmission of high bandwidth multimedia data, including advanced 3D video applications. Such wireless multimedia systems should be designed with the purpose of maximizing the quality perceived by the users. For instance, quality parameters can be measured at the receiver-side and fed back to the transmitter for system optimization. Measuring 3D video quality is a challenge due to a number of perceptual attributes associated with 3D video viewing (e.g., image quality, depth perception, naturalness). Subjective as well as objective metrics have been developed to measure 3D video quality against different artifacts. However most of these metrics are Full-Reference (FR) quality metrics and require the original 3D video sequence to measure the quality at the receiver-end. Therefore, these are not a viable solution for system monitoring/update "on the fly." This chapter presents a Near No-Reference (NR) quality metric for color plus depth 3D video compression and transmission using the extracted edge information of color images and depth maps. This work is motivated by the fact that the edges/contours of the depth map and of the corresponding color image can represent different depth levels and identify image objects/boundaries of the corresponding color image and hence can be used in quality evaluation. The performance of the proposed method is evaluated for different compression ratios and network conditions. The results obtained match well those achieved with its counterpart FR quality metric and with subjective tests, with only a few bytes of overhead for the original 3D image sequence as side-information.