CCSF-Quality: Defining new Chromatic Contrast Sensitivity Functions for improved quality assessment and quality enhancement

The Human Visual System (HVS) has long been an inter-disciplinary field of research for different scientific communities such as psychologists, cognitive scientists, and neuroscientists but also researchers in fields such as computer vision, image processing, and color science. Among different topics addressed in research on the HVS, dealing with different aspects of contrast in an image (such as evaluation, enhancement, and modeling) especially with regards to chromatic contrast is of high interest for different research groups. Possible applications of such research especially in the field of computer vision and image processing are objective image and video quality assessment, image and video compression, gamut mapping, and halftoning. In all such applications researchers try to model the HVS in the best possible way to provide an accurate method and approach. After an in-depth literature review on previous works in this field, the subjective experiment was carefully planned. Starting from February 2018 observers were asked to participate in the test and data collection was started. The experimental setup is now optimized and so a higher number of observation data has been collected in the dataset compared to any other published work. While we have already published initial report on the dataset at CIC 2019 a more detailed work is in preparation and will soon be submitted to a high impact journal. Apart from performing the subjective experiment, a CNN based image quality metric was proposed and presented at the Color Imaging Conference (CIC) 2018. Two other publications on contrast evaluation and enhancement methods was also published at EUVIP 2019 which was the result of a work by a master student co-supervised by the project leader. During the last year, the project leader Dr. Seyed Ali Amirshahi who was based at the Laboratoire de Traitement et Transport de l'Information at University of Paris 13 for the first two years of the project moved back to the Norwegian Colour and Visual Computing Laboratory at Norwegian University of Science and Technology (NTNU) in Gjøvik. It should also be pointed out that after the end of the project, Dr. Amirshahi has been employed as an associated professor at NTNU.

The finding of the project would help researchers to better model the human visual system and allow the introduction of new image and video quality metrics and enhancement methods, image and video compression, gamut mapping and halftoning. The subjective dataset is the largest of such datasets in the field and it will soon be publicly released. We anticipate that this dataset will be an important resource for the scientific community. The project has further strengthened the collaboration with University Paris 13. The Project Leader (PL) also attended the Huawei Future Device Technology Summit at Helsinki, Finland as an invited speaker. At the end of the project the PL has been employed as an associate professor at NTNU. This is in line with the goal of the FRIPRO Mobility Grant which is to train researchers for the Norwegian research community.

The Human Visual System (HVS) has long been an inter-disciplinary field of research for different scientific communities such as psychologists, cognitive scientists, and neuroscientists but also researchers in fields such as computer vision, image processing, and color science. Among different topics addressed in research on the HVS, dealing with different aspects of contrast in an image (such as evaluation, enhancement, and modeling) especially with regards to chromatic contrast is of high interest for different research groups. Possible applications of such research especially in the field of computer vision and image processing are objective image and video quality assessment, image and video compression, gamut mapping, and halftoning. In all such applications researchers try to model the HVS in the best possible way to provide an accurate method and approach. Through subjective tests, this project aims to first find a new Chromatic Contrast Sensitivity Function (CCSF) to model observer's sensitivity towards changes in the chromatic contrast of an image. Using the introduced function, new Image Quality and Video Quality Metrics will be introduced. With the huge amount of images and videos in our daily life there has been a significant increase in research on metrics which are able to evaluate their quality. Finally, work on image and video quality enhancement will be performed taking advantage of the new image and video quality metrics and using the new CCSF. The project which will be managed by Dr. Seyed Ali Amirshahi includes two prestigious national and international groups, the Norwegian Colour and Visual Computing Laboratory at Gjøvik University College and the Laboratoire de Traitement et Transport de l'Information at University of Paris 13. Being awarded such opportunity will most definitely benefit the project manager (Dr. Amirshahi) who as a young and less experienced researcher plans to continue his career in academia and especially in Norway.