Light Field Image And Preprocessed-transcoded Video Quality Assessment

With the rapid development of media technology and the popularity of mobile dis-play devices,there are also higher expectations for social media messages.On the one hand,people expect to receive media information with a better sense of telepresence and immersion.On the other hand,they also want to be the providers of images and videos,rather than the recipients only.Therefore,many emerging media modalities have been proposed and applied,such as Light Field Images(LFI)and streaming videos.How-ever,processing such as encoding and transmission will degrade the perceptual quality of images and videos.It is very necessary for the image/video quality assessment.This work focuses on the two core issues:light field image quality assessment and streaming video quality assessment,and carries out deep research on these issues.The LFI quality depends on both angular consistency and spatial quality.However,few existing LFI quality assessment methods concentrate on effects caused by angular inconsistency.Since the Micro-Lens Image(MLI)refers to the angular domain of the LFI,which can simultaneously record the angular information in both horizontal and vertical directions,we propose a No-Reference Light Field image Quality assessment model based on MLI(LF-QMLI).Specifically,we first utilize Global Entropy Distri-bution(GED)and Uniform Local Binary Pattern descriptor(ULBP)to extract features from the MLI and then pool them together to measure angular consistency.Also,the information entropy of Sub-Aperture Image(SAI)is adopted to measure spatial qual-ity.Extensive experimental results on public databases show that LF-QMLI achieves state-of-the-art performance.On the other hand,the pre-processed video transcoding has attracted wide atten-tion and has been increasingly used in practical applications for improving the per-ceptual experience and saving transmission resources.In this paper,we select the source videos and various pre-processing approaches to construct the first Pre-processed and Transcoded Video Database(PTVD).Then,we conduct the subjective experiment,showing that compared with the video sent to the codec directly at the same bitrate,the appropriate pre-processing methods indeed improve the perceptual quality.Finally,existing image/video quality metrics are evaluated on our database.The results indi-cate that the performance of the existing image/video quality assessment approaches remains to be improved.Then we propose a video quality assessment model based on the human visual tem-poral hysteresis effect for pre-processed video transcoding application scenarios and it can meet real-time requirements.First,we evaluate the quality of each video frame through the single-frame model and consider the correction of perceptual quality by the pre-processing algorithm.Then the overall quality of the video is predicted by com-bining all video frame values with the temporal pooling model.Experimental results show that the performance of our proposed model significantly outperforms the existing objective quality assessment algorithms,and meets the real-time requirements.