—----------------------------------------------------------------------------------------------------------------------------VSQuAD - Video Surveillance Quality Assessment Dataset Copyright(c) 2022 - All Rights Reserved. —---------------------------------------------------------------------------------------------------------------------------- Permission to use, copy, or modify this database and its documentation for educational and research purposes only and without fee is hereby granted, provided that this copyright notice and the original authors' names appear on all copies and supporting documentation. This database shall not be used, redistributed, or adapted as the basis of a commercial software or hardware product without first obtaining permission of the authors. The authors make no representations about the suitability of this database for any purpose. It is provided "as is" without express or implied warranty. ------------------------------------------------------------------------------------------------------------------------------- Investigators: The main investigators of the work are: * Azeddine Beghdadi, Mounir Kaaniche, Borhen-eddine Dakkar ; L2TI lab, University Sorbonne Paris Nord, France * Muhammad Ali Qureshi, Hammad Hassan Gillani; The Islamia University of Bahawalpur (Pakistan), * Zohaib Amjad Khan; L2S, CentraleSupélec, University Paris Saclay, France * Faouzi Alaya Cheikh, Mohib Ullah; Norwegian University of Science and Technology, Norway, How to cite the database: The associated paper containing more details and analysis will be submitted to ICIP 2022 for consideration. Contacts: * Prof. Azeddine Beghdadi, L2TI, Université Sorbonne Paris Nord, France beghdadi@sorbonne-paris-nord.fr * Muhammad Ali Qureshi, The Islamia University of Bahawalpur, Pakistan ali.qureshi@iub.edu.pk * Borhen-eddine Dakkar, Université Sorbonne Paris Nord, France dakkarborheneddine@gmail.com —---------------------------------------------------------------------------------------------------------------------------- Below are the details on the VSQuAD dataset —---------------------------------------------------------------------------------------------------------------------------- Overview The VQUAD dataset is specifically built with the aim for video quality assessment of surveillance videos. The database consists of Thirty-six (36) reference videos with a full HD resolution of 1920 x 1080 in mp4 format. The videos are of 10 seconds duration with frame rates of 30 frames per second. Each of the videos has been distorted by nine (09) different distortions at four (04) different levels. The database also contains mixtures of distortions formed by some of the common degradations for particular scenarios. This resulted in a total of 1576 distorted videos. The dataset contains different scenarios corresponding to the most common places where video surveillance systems are the most used. The scenarios used in the dataset are: Airport, Car Moving on Bridge, Bus by night, Monument place, Harbor, Landing plane, Highway, Bus Stop, Traffic Road, Manifestation, Mountain village, Outside Bank, Outside Shop, Parking ,Pedestrian, Railway Platform, River View, Roma by night, Saida night, Square place, Stadium, Tour de France, Tram Station, Vatican square night, boat on river, car street by night, mall, School Nine different types of video distortions are introduced: - DB (Defocus Blur), HZ (Haze), LI (Low-Illumination), MB (Motion Blur), WN (AWGN), RN (Rain), SK (Smoke), UI (Uneven Illumination), CA (Compression Artifacts). For each source video, 4 different levels were applied in the case of single distortion. In the case of multiple distortions (2 or 3) applied simultaneously, 4 cases of combinations of severity levels are used for each mixture. Single Distortions: D1 Defocus blur (DB) D2 Haze (HZ) D3 Low illumination (LI) D4 Motion blur (MB) D5 AWGN (WN) D6 Rain (RN ) D7 Smoke (SM) D8 Uneven illumination (UI) D9 Compression artifacts (CA) Distortion Levels: 1 Just Noticeable / Hardly visible 2 Visible but Not Annoying 3 Annoying 4 Very Annoying —---------------------------------------------------- Multiple Distortions: M1 D1-D2 M2 D3-D5 M3 D5-D8 M4 D3-D6 M5 D1-D9 M6 D3-D5-D6 M7 D5-D8-D9 M8 D4-D6-D8 M9 D2-D5-D9 Distortion Levels: In the case of multiple distortions, 4 combinations of the severity levels, among all possibilities, are used. Folder "Reference videos": This folder provides all 36 reference videos. Folder "Distorted videos": This folder provides 9 subfolders of single distortion: Defocus blur, Haze, Low light, Motion blur, Smoke, Uneven illumination, WN, Compression, Rain, ...and 9 subfolders of multi distortions: M1, M2, M3, M4, M5, M6, M7, M8, M9. For each source video, 36 videos (corresponding to 9 distortion x 4 levels), while all 36 possible combinations of these video files constitute the distortion sets of each source content. File Naming Conventions: File names are organized in such a manner that they indicate the name of the reference video, followed by the distortion type label "Dj", then a number of distortion's level k. i.e. "videoName_Dj_k.mp4". Example of file name Airport_D1_2.mp4 Airport video degraded with distortion D1 (Defocus Blur) at level 2. Airport_D3_D1.mp4 Airport video degraded with mixture M1 consisting of distortions D3 (Low illumination) and D1 (Defocus Blur). BusStop_D3_D5_D6.mp4 BusStop video degraded with mixture M6 consisting of distortions D3 (Low illumination), D5 (AWGN) and D6 (Rain) . Note that in the case of mixture we are only interested in the identification of the different distortions and not in their level of severity. ______________________________ END _________________________________________