2026-05-22T14:07:51Zhttps://keep.lib.asu.edu/oai/requestoai:keep.lib.asu.edu:node-1563842024-12-20T18:25:12Zoai_pmh:alloai_pmh:repo_items156384
https://hdl.handle.net/2286/R.I.49241
http://rightsstatements.org/vocab/InC/1.0/
All Rights Reserved
2018
116 pages
Doctoral Dissertation
Academic theses
Text
eng
Golestaneh, Seyedalireza
Karam, Lina
Bliss, Daniel W.
Li, Baoxin
Turaga, Pavan K.
Arizona State University
Doctoral Dissertation Electrical Engineering 2018
Digital imaging and image processing technologies have revolutionized the way in which<br/><br/>we capture, store, receive, view, utilize, and share images. In image-based applications,<br/><br/>through different processing stages (e.g., acquisition, compression, and transmission), images<br/><br/>are subjected to different types of distortions which degrade their visual quality. Image<br/><br/>Quality Assessment (IQA) attempts to use computational models to automatically evaluate<br/><br/>and estimate the image quality in accordance with subjective evaluations. Moreover, with<br/><br/>the fast development of computer vision techniques, it is important in practice to extract<br/><br/>and understand the information contained in blurred images or regions.<br/><br/>The work in this dissertation focuses on reduced-reference visual quality assessment of<br/><br/>images and textures, as well as perceptual-based spatially-varying blur detection.<br/><br/>A training-free low-cost Reduced-Reference IQA (RRIQA) method is proposed. The<br/><br/>proposed method requires a very small number of reduced-reference (RR) features. Extensive<br/><br/>experiments performed on different benchmark databases demonstrate that the proposed<br/><br/>RRIQA method, delivers highly competitive performance as compared with the<br/><br/>state-of-the-art RRIQA models for both natural and texture images.<br/><br/>In the context of texture, the effect of texture granularity on the quality of synthesized<br/><br/>textures is studied. Moreover, two RR objective visual quality assessment methods that<br/><br/>quantify the perceived quality of synthesized textures are proposed. Performance evaluations<br/><br/>on two synthesized texture databases demonstrate that the proposed RR metrics outperforms<br/><br/>full-reference (FR), no-reference (NR), and RR state-of-the-art quality metrics in<br/><br/>predicting the perceived visual quality of the synthesized textures.<br/><br/>Last but not least, an effective approach to address the spatially-varying blur detection<br/><br/>problem from a single image without requiring any knowledge about the blur type, level,<br/><br/>or camera settings is proposed. The evaluations of the proposed approach on a diverse<br/><br/>sets of blurry images with different blur types, levels, and content demonstrate that the<br/><br/>proposed algorithm performs favorably against the state-of-the-art methods qualitatively<br/><br/>and quantitatively.
Engineering
Visual Quality Assessment and Blur Detection Based on the Transform of Gradient Magnitudes