Blind Image Quality Assessment Based on Multi-scale KLT

Blind image quality assessment (BIQA) plays an important role in image services as independent of the reference image. Herein, the perceptual relevant feature design is the core of BIQA methods, but their performance is still not satisfied at present. In this work, we propose an unsupervised feature extraction approach for BIQA based on Karhunen-Loeve transform (KLT). Specifically, a normalization operation is firstly applied to the test image by calculating its mean subtracted contrast normalized (MSCN) coefficient. Then, KLT is employed as a data-driven feature extraction approach to extract image structural features, wherein kernels with different sizes are utilized to perform multi-scale analysis. Finally, generalized Gaussian distribution (GGD) is employed to model the KLT coefficients distribution in different spectral components as quality relevant features. Extensive experiments conducted on four widely utilized IQA databases have demonstrated that the proposed Multi-scale KLT (MsKLT) BIQA metric compares favorably with existing BIQA methods in terms of high accordance with human subjective scores on both common and uncommon distortion types.

Automated summary

In this work, an unsupervised feature extraction approach based on Karhunen-Loeve transform (KLT) for blind image quality assessment (BIQA) is proposed. A normalization operation and KLT are used to extract image structural features, with generalized Gaussian distribution employed to model the KLT coefficients distribution as quality relevant features. Extensive experiments show that the proposed method outperforms existing BIQA methods in terms of agreement with human subjective scores on various types of distortions.