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.

No abstract

Background: QT interval (QT) genome-wide association studies (GWAS) have identified upwards of 35 common variant loci, including SNPs adjacent to putative transcription elongation factor TCEA3. Transcription elongation control has broad effects on gene expression, the misregulation of which is known to influence cardiac conduction system morphogenesis as well as activation or repression of key regulatory genes. Thus, we hypothesized that a genome-wide gene-gene interaction study of TCEA3 lead SNP rs2298632 would identify novel loci that influence QT. Methods: Using 1000 Genomes imputed data (>20 million SNPs) in n=67,445 participants (69% Caucasian; 18% Hispanic/Latino; 11% African American) from 10 studies, we conducted genome-wide meta-analyses to test for the presence of: interaction effect loci by examining rs2298632xSNP interactions on QT; and joint effect loci by simultaneously examining SNP main effects and rs2298632xSNP interactions on QT. Inverse-variance weighted meta-analysis of genomically controlled ancestry- and study-specific summary effects estimated using multivariable adjusted linear models or generalized estimating equations that incorporated robust standard errors was performed using METAL. SNPs demonstrating evidence of heterogeneity (Cochran’s Q P < 0.05) and SNPs that were infrequent or rare (minor allele frequency [MAF] <5%) were excluded. Results: We identified one genome-wide significant interaction effect locus ( P INT <5x10 -8 ) at PVT1 (lead SNP: rs4733591; mean MAF = 32%), a long non-coding RNA gene for which previous GWAS identified suggestive associations with left ventricular systolic dysfunction. We also identified four genome-wide significant joint effect loci ( P JOINT <5x10 -8 ) that mapped within or nearby NUCKS1 (lead SNP: rs823094; MAF = 0.33) , CASR (lead SNP = rs17251221; MAF = 0.13) , ACTBL2 (lead SNP = rs7737409; MAF = 0.22) , and KDM1B (lead SNP = rs34969716; MAF = 0.26) , loci with roles in calcium sensing, regulation of gene expression through histone modification, and tumor suppression. Conclusion: Extension of traditional main effects GWAS to interrogate gene-gene interactions for biologically motivated loci like TCEA3 may help inform the structure and function of genetic pathways underlying complex traits like QT.