Screen content image quality assessment and processing

• Main Author: Guo, Heng
• Other Authors: Kai-Kuang Ma
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/149257
• Institution: Nanyang Technological University

With rapid development of information technology, screen content images (SCIs), which contain a mixture of natural images together with texts, charts, symbols, and/or computer generated graphics, are often encountered in various multimedia applications. SCIs usually contain a large amount of sharp edges, hence the perception of SCIs are quite different from that of natural images. However, how to display the SCIs with a high quality remains an unmet need in image processing, and even the state-of-the-art technology is not able to handle SCIs well considering the extensive amount of sharp edges in these images. To address this problem, three fundamental image processing techniques for screen content color images are investigated in this thesis; i.e., image quality assessment, image blurriness estimation, and image lightness estimation. Among these techniques, image quality assessment is a fundamental tool to objectively evaluate the quality of the screen content color image from the observer's point of view. Image blurriness estimation distinguishes the blurred regions and sharp regions of an image, providing a fundamental guidance to image enhancement algorithms to improve the perceptual quality of images. Finally, the image lightness estimation method is a generic tool for extracting the lightness component from an image, which is able to help facilitate a variety of image processing tasks that manipulate the lightness of images.