Blind Stereoscopic Image Quality Assessment Based on Hierarchical Learning

Abstract

We proposed a blind image quality assessment model which used classification and prediction for three-dimensional (3D) image quality assessment (denoted as CAP-3DIQA) that can automatically evaluate the quality of stereoscopic images. First, in the classification stage, the model separated the distorted images into several subsets according to the types of image distortions. This process will assign the images with the same distortion type to the same group. After the classification stage, the classified distorted image set is fed into the image quality predictor that contains five different perceptual channels which predict the image quality score individually. Finally, we used the regression module of the support vector machine to evaluate the final image quality score, where the input of the regression model is the combination of five channel's outputs. The model, we proposed is tested on three public and popular databases, which are LIVE 3D Image Quality Database Phase I, LIVE 3D Image Quality Database Phase II, and MCL 3D Image Quality Database. The experimental results show that our proposed model leads to significant performance improvement on quality prediction for stereoscopic images compared with other existing state-of-the-art quality metrics.INDEX TERMS Hierarchical learning, image quality assessment, no reference, stereoscopic images.

Blind Stereoscopic Image Quality Assessment Accounting for Human Monocular Visual Properties and Binocular Interactions

“…We compare the proposed model with three 2D IQA models (SSIM, BRISQUE [62] and ADD-GSIM [72]) and eleven 3D metrics: Shao2016 [16], Shao2017 [73], Zhou2017 [10], Zhou'2017 [12], Liu2018 [74], Wang2018 [75], Ma2018 [20], Yue2018 [40] Shao2018 [17], Chen2019 [21], and Liu2019 [18], on LIVE 3D Phase I and LIVE 3D Phase II. Works conducted on MCL 3D Database are relatively fewer, so here we take eight works compared with our proposed model on MCL 3D Database: SSIM, BRISQUE, Shao2016, Zhou2017, Liu2018 [74], Yang'2018 [15], Chen2019 and Liu2019.…”

“…To handle the asymmetric distortion, Shao et al [17] designed an NR metric by utilizing both monocular and binocular properties for quality assessment. Likewise, Liu et al [18] combined the monocular and binocular information and proposed a blind SIQA model by using the SVM method to fuse the quality score.…”

Blind Stereo Image Quality Evaluation Based on Convolutional Network and Saliency Weighting

“…With the rapid development of stereo image applications, many related stereo image technologies and services have been introduced in our daily lives as well as in many professional fields [1][2][3][4][5][6][7][8][9]. A variety of distortions can occur during the collection, transmission, processing, and displaying of stereo images [10][11][12][13][14][15][16][17][18][19]. erefore, it is of immense practical significance to establish a high-performance stereo image quality evaluation method.…”

“…erefore, objective evaluation plays a dominant role in SIQE. ere are three main types of objective stereo image quality evaluation methods: full-reference (FR) evaluation [4][5][6][7][8], reduced-reference (RR) evaluation [9], and no-reference (NR) evaluation [10][11][12][13][14][15][16][17][18][19][20][21][22]. e FR evaluation method compares the undistorted original image with the distorted image to obtain the difference between them.…”

“…Wu et al [17] proposed an evaluation method based on depth edge information and color signals, which also uses segmented self-encoders. Liu et al [18] proposed an SIQE method based on classification and prediction. Jiang et al proposed an SIQE algorithm for processing multiple distortions [19], which characterizes the local receptive field characteristics of the visual cortex by learning monocular and binocular local visual primitives for various distorted stereo images quality assessment and single distortion stereo images.…”