----------------- SIAT Synthesized Video Quality Database ----------------- Yun Zhang - Email:yun.zhang@siat.ac.cn --------- Reference --------- 1. X. Liu, Y. Zhang, S. Hu, S. Kwong, C.-C. J. Kuo, and Q. Peng, "Subjective and Objective Video Quality Assessment of 3-D Synthesized View with Texture/Depth Compression Distortion",IEEE Transactions on Image Processing, vol.24, no.12, pp.4847-4861, Dec. 2015. 2. URL: http://codec.siat.ac.cn/SIAT_Synthesized_Video_Quality_Database/index.html ------- Sequences ------- "Sequences" provides sequences in the SIAT Synthesized Video Quality Database which are YUV 4:2:0 format. There are ten MVD sequences used in this database and the information are as follows: BookArrival, 1024x768, 16.67fps, 100 frames, synthesized view: 8 (from view 6-10) Balloons, 1024x768, 30fps, 200 frames, sythesized view: 3 (from view 1-5) Kendo, 1024x768, 30fps, 200 frames, synthesized view: 3 (from view 1-5) lovebird1, 1024x768, 30fps, 200 frames, synthesized view: 5 (from view 4-6) Newspaper, 1024x768, 30fps, 200 frames, synthesized view: 3 (from view 2-4) Dancer, 1920x1088, 25fps, 200 frames, synthesized view: 5 (from view 1-9) PoznanHall2, 1920x1088, 25fps, 200 frames, synthesized view: 6 (from view 5-7) PoznanStreet, 1920x1088, 25fps, 200 frames, synthesized view: 4 (from view 3-5) GT Fly, 1920x1088, 25fps, 200 framses, synthesized view: 5 (from view 1-9) Shark, 1920x1088, 25fps, 200 framses, synthesized view: 5 (from view 1-9) The Dancer, GT Fly and Shark sequences are computer animation videos, and the others are natural scenes. There exist camera movement in sequences Balloons, Kendo, Dancer, GT Fly, Shark and PoznanHall2, while the cameras in BookArrival, Lovebird1, Newspaper, PoznanStreet are still. ------------------------ File Naming Conventions ------------------------ There are a total of 150 videos in the folder, 15 videos for each sequence. An example of the file naming covention is as follows: 1. "Balloons_1Dfast.yuv" is the video synthesized with VSRS-1D-Fast software from the original texture video and the original depth video. 2. "Balloons_1Dfast_depthQP32.yuv" is the video synthesized with VSRS-1D-Fast software from the original texture video and the compressed depth video (compression quatization parameter is 32). 3. "Balloons_1Dfast_textureQP24.yuv" is the video synthesized with VSRS-1D-Fast software from the compressed texture video (compression quatization parameter is 24) and the original depth video. 4. "Balloons_1Dfast_textureQP24_depthQP32.yuv" is the video synthesized with VSRS-1D-Fast software from the compressed texture video (compression quatization parameter is 24) and the compressed depth video (compression quatization parameter is 32). 5. "Balloons_ref.yuv" is the original reference video in the virtual view position where the synthesized video is generated. --------------- Subjective Data --------------- The experiment was conducted in two sessions. 84 videos were evaluated in the first session and 56 videos were evaluated in the second session. There were a total of 56 non-expert subjects (38 males and 18 females with an average age of 24) participated in the experiment and each test video was evaluated by 40 subjects. The quality scores assigned by each subject should be normalized per person before computing DMOS. Because there were 24 subjects participated in both the first and the second sessions, to avoid inconsistent scoring criteria of the same subject between two sessions, all the reference videos were included in both sessions using hidden reference removal procedure. The DMOS score for each test video was computed per session using the quality score assigned to the corresponding reference video in that session. All subjects taking part in the experiment are graduate students in the school of information science and technology. Every subject had been told the procedure of the test and had watched a training sequence before starting the experiment. After finishing the rating process, subjects with extreme scores should be rejected as recommended in ITU-R.BT500. In this experiment, only 1 subject (male) in the first session was eliminated, and there were 55 remaining subjects in our final database. Then, difference scores were obtained by subtracting the score of the test sequence by the score of the reference sequence. Both scores were given by the same subject in the same session. There are 4.13% (228 of 5516) difference scores which are negative, i.e., the score of the reference video was less than the score of the tested video. These negative difference scores might come from the special preference of some subjects, but not from experiment mistakes. We keep these scores negative to ensure the original diversity of the subjective data. The difference scores are then normalized to z-scores per person and per session. To make the data more intuitive, the normalized z-scores were then linearly re-scaled to a range of [0,1]. Finally, we computed DMOS for each synthesized video. Note: To obtain the DMOS score, please contact us via the following email: Yun Zhang - Email:yun.zhang@siat.ac.cn