• Optics and Precision Engineering
  • Vol. 33, Issue 1, 107 (2025)
Jiaxin LI, Fajie DUAN*, Xiao FU, and Guangyue NIU
Author Affiliations
  • State Key Laboratory of Precision Measuring Technology & Instruments, Tianjin University, Tianjin300072, China
  • show less
    DOI: 10.37188/OPE.20253301.0107 Cite this Article
    Jiaxin LI, Fajie DUAN, Xiao FU, Guangyue NIU. Full-reference image quality assessment based on texture singular value decomposition[J]. Optics and Precision Engineering, 2025, 33(1): 107 Copy Citation Text show less

    Abstract

    In industrial vision systems, subjective assessment is costly, pre-training for no-reference quality evaluation is time-intensive, and there is a critical need for a highly accurate full-reference image quality assessment model. To address these challenges, this study proposes a novel full-reference image quality assessment model based on singular value decomposition (SVD) with weighted texture information. First, SVD is applied to the reference image blocks, and the singular values of the distorted blocks are estimated using the singular vectors of both the reference and distorted image blocks, yielding the brightness similarity component. Next, the estimated singular values of the distorted image blocks are used to quantify average offset distortion and contrast change distortion, resulting in the contrast similarity component. The structural similarity of the images is then determined by analyzing the deviation of the singular vectors of the distorted image blocks from the unit matrix of the reference image blocks. Finally, the brightness, contrast, and structural similarity components are weighted using texture information to construct the full-reference image quality assessment model. The proposed method was evaluated on six widely used image quality assessment databases across four performance criteria. Experimental results demonstrate that the model achieves a weighted Spearman rank correlation coefficient of 0.896 3 across the datasets. For contrast change distortion, the model attains a Spearman rank correlation coefficient of 0.859 5, outperforming the second-best method by 85%. Compared to existing full-reference image quality assessment models, the proposed approach offers significant advantages in prediction accuracy, generalization capability, and computational efficiency.
    Am×n=i=1kSi×Ui×ViT k=min(m,n)(1)

    View in Article

    Am×n=Um×mSm×nVn×nT(2)

    View in Article

    S^i=A^Vi      i=1,2,,mU^i=0if   s^i=0A^Vi/S^iotherwise(3)

    View in Article

    FLU=i=1ThrSi-S^LUiwi,FLV=i=1ThrSi-S^LViwi,(4)

    View in Article

    wi=Si/i=1block sizeSi+c(5)

    View in Article

    i=1ThrSi-S^LUi+Si-S^LVii=1NOSSi-S^LUi+Si-S^LVi=δ(6)

    View in Article

    FL=eFLU+FLV22(7)

    View in Article

    FCm=e-S1-S1Tu×S^12(8)

    View in Article

    Dvs=S^S(9)

    View in Article

    FCc=i=2Thrmin(γ,Dvsi)×wi(10)

    View in Article

    FC=FCm×FCc(11)

    View in Article

    FSU=i=1ThrSUi-1wi2,FSV=i=1ThrSVi-1wi2,(12)

    View in Article

    FS=e-FSU+FSV2(13)

    View in Article

    S=FLαFCβFSγ(14)

    View in Article

    Gn(x,y)=Gx2(x,y)+Gy2(x,y)(15)

    View in Article

    Gh=i=1block sizeGn2(x,y)(16)

    View in Article

    δi=Ghii=1#blocksGhi(17)

    View in Article

    TSVD=i=1#blocksSi×δi(18)

    View in Article

    PLCC=i=1nsi-s¯qi-q¯i=1nsi-s¯2i=1nqi-q¯2(19)

    View in Article

    SROCC=1-6i=1ndi2n2n2-1(20)

    View in Article

    KROCC=2Nc-NdN(N-1)(21)

    View in Article

    f(s)=β112-11+expβ2s-β3+β4s+β5(22)

    View in Article

    Jiaxin LI, Fajie DUAN, Xiao FU, Guangyue NIU. Full-reference image quality assessment based on texture singular value decomposition[J]. Optics and Precision Engineering, 2025, 33(1): 107
    Download Citation