Open AccessDeep CNN Model for Non-Screen Content and Screen Content Image Quality Assessment
Author(s) -
Dilip Chaudhary,
Venkatesh Venkatesh
Publication year - 2022
Publication title -
global journal of computer science and technology
Language(s) - English
Resource type - Journals
ISSN - 0975-4172
DOI - 10.34257/gjcstdvol22is1pg17
Subject(s) - computer science , content (measure theory) , image (mathematics) , quality (philosophy) , image quality , multimedia , user generated content , work (physics) , artificial intelligence , computer vision , world wide web , social media , mathematical analysis , philosophy , mathematics , epistemology , mechanical engineering , engineering
In the current world, user experience in various platforms matters a lot for different organizations. But providing a better experience can be challenging if the multimedia content on online platforms is having different kinds of distortions which impact the overall experience of the user. There can be various reasons behind distortions such as compression or minimal lighting condition while taking photos. In this work, a deep CNN-based Non-Screen Content and Screen Content NR-IQA framework is proposed which solves this issue in a more effective way. The framework is known as DNSSCIQ. Two different architectures are proposed based upon the input image type whether the input is a screen content or non-screen content image. This work attempts to solve this by evaluating the quality of such images.