Title :
Image quality assessment based on gradient complex matrix
Author_Institution :
Key Lab. of Airborne Opt. Imaging & Meas., Changchun Inst. of Opt., Fine Mech. & Phys., Changchun, China
Abstract :
An image quality assessment model based on gradient complex matrix is proposed. The vertical and horizontal gradient information of grayscale image is calculated. Complex number is used to construct the measuring matrix. Singular value decomposition is performed in order to obtain the main structure information of the image. The singular value feature vectors of the image gradient complex matrices corresponding to the reference image and the distorted image are used to measure the structural similarity of the two images. PSNR is taken as a tool to evaluate the gradient distribution similarity. Their properties are analyzed by using LIVE database and nonlinearity regression function.
Keywords :
feature extraction; gradient methods; image matching; matrix algebra; regression analysis; singular value decomposition; visual databases; LIVE database; PSNR; complex number; feature vectors; gradient complex matrix-based image quality assessment; gradient distribution similarity; grayscale image; horizontal gradient information; image distortion; image gradient complex matrices; image similarity; image structure information; measuring matrix; nonlinearity regression function; reference image; singular value decomposition; vertical gradient information; Distortion measurement; Gray-scale; Humans; Image quality; Matrix decomposition; PSNR; Periodic structures; complex matrix; gradient; image quality assessment; singular value decomposition;
Conference_Titel :
Systems and Informatics (ICSAI), 2012 International Conference on
Conference_Location :
Yantai
Print_ISBN :
978-1-4673-0198-5
DOI :
10.1109/ICSAI.2012.6223427
