Image Retargeting Subjective Quality Database

Lin Ma	Weisi Lin	Chenwei Deng	King Ngi Ngan
lma@ee.cuhk.edu.hk	wslin@ntu.edu.sg	cwdeng@ntu.edu.sg	knngan@ee.cuhk.edu.hk
Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong	School of Computer Engineering, Nanyang Technological University, Singapore	School of Computer Engineering, Nanyang Technological University, Singapore	Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong

Source Image Attribute Information:

We have considered four attributes, specifically face and people, clear foreground object, natural scenery, and geometric structure. We define the clear foreground object attribute as that the salient object should occupy the image region smaller than 50% of the source image. The natural scenery attribute means that a large proportion of the image contains the texture or smooth information. And the geometric structure attribute denotes that there are evident edges or lines in the source image. The detailed attribute information of each source image is illustrated in the following table. Firstly, it can be observed that one image may contain more than one attributes. Secondly, the dominant attribute of each source image is illustrated. We sort the attributes of each source image according to the attribute importance.

The attribute information of the source image. (1 indicates the attribute of “face and people”; 2 indicates the attribute of “clear foreground object”; 3 indicates the attribute of “natural scenery”; 4 indicates the attribute of the “geometric structure”. The attributes are sorted according to their corresponding importance. The left attribute denotes the most important, while the right one is the least important.)

Detailed information about the source image attribute information can be found in the SUPPLEMENTARY.

Retargeted Image and the Obtained MOS Value:

The obtained MOS value of each retargeted image after the subjective testing and data processing. The horizontal axes corresponds to the image number, and the vertical axes corresponds to the MOS value. The red star indicates the obtained MOS value for each retargeted image. The blue error bar indicates the standard deviation of the subjective scores.

There are in total 171 retargeted images in the built database. Detailed information about the retargeted image name and its corresponding number are shown in the SUPPLEMENTARY. It can be observed that each source image shown in session 1 generates 3 retargeted images by different methods and in different scales.

Histogram of the MOS values in 15 equally spaced bins between the minimum and maximum MOS values of the image retargeting database (In total, 171 retargeted images are included in the database).

The perceptual qualities of the retargeted images in the database should span the entire range of visual quality and exhibit good perceptual quality separation. The histogram of the MOS values is illustrated above. It can be observed that the perceptual qualities of the images range from low to high values. Also it demonstrates that the subjective study samples a range of perceptual quality in an approximately uniform fashion. The image perceptual qualities exhibit a good separation.

Copyright Notice:

All rights reserved.Permission is hereby granted, without written agreement and without license or royalty fees, to use, copy, and distribute this database (the images, the results, and the source files) and its documentation for non-commercial research and educational purposes only, provided that the copyright notice in its entirety appear in all copies of this database, and the original source of this database, Image and Video Processing Laboratory (IVP, http://ivp.ee.cuhk.edu.hk/) at The Chinese University of Hong Kong, is acknowledged in any publication that reports research using this database.

IN NO EVENT SHALL THE CHINESE UNIVERSITY OF HONG KONG BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL LOSS OR DAMAGES ARISING OUT OF OR IN CONNECTION WITH THE USE OF OR THE INABILITY TO USE THIS DATABASE, ITS DOCUMENTATION AND INFORMATION, EVEN IF THE CHINESE UNIVERSITY OF HONG KONG HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH LOSS OR DAMAGE.

THE CHINESE UNIVERSITY OF HONG KONG SPECIFICALLY DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, ACCURACY OR COMPLETENESS AND NON INFRINGEMENT OF INTELLECTUAL PROPERTY RIGHTS OF ANY PARTY. THE DATABASE PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND THE CHINESE UNIVERSITY OF HONG KONG HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.