Research Article
Saliency Detection Using Sparse and Nonlinear Feature Representation
Table 4
Comparison of AUC and PoDSC score of the proposed model with other state-of-the-art models.
| Algorithm | C1 | C2 | C3 | C4 | C5 | C6 | Avg. | AUC | PoDSC | AUC | PoDSC | AUC | PoDSC | AUC | PoDSC | AUC | PoDSC | AUC | PoDSC | Avg. AUC |
| AIM | 0.921 | 0.653 | 0.918 | 0.543 | 0.957 | 0.461 | 0.904 | 0.435 | 0.924 | 0.508 | 0.941 | 0.671 | 0.927 | MESR | 0.798 | 0.494 | 0.854 | 0.436 | 0.927 | 0.377 | 0.718 | 0.238 | 0.837 | 0.398 | 0.918 | 0.617 | 0.842 | ICL | 0.928 | 0.681 | 0.907 | 0.501 | 0.909 | 0.348 | 0.926 | 0.537 | 0.902 | 0.498 | 0.914 | 0.569 | 0.914 | Itti | 0.939 | 0.687 | 0.924 | 0.542 | 0.952 | 0.457 | 0.882 | 0.423 | 0.924 | 0.499 | 0.941 | 0.631 | 0.927 | MPQFT | 0.805 | 0.497 | 0.865 | 0.452 | 0.927 | 0.386 | 0.735 | 0.248 | 0.852 | 0.436 | 0.921 | 0.630 | 0.851 | SDSR | 0.879 | 0.618 | 0.908 | 0.511 | 0.948 | 0.435 | 0.801 | 0.298 | 0.904 | 0.489 | 0.933 | 0.645 | 0.896 | SUN | 0.766 | 0.459 | 0.787 | 0.349 | 0.880 | 0.362 | 0.708 | 0.263 | 0.719 | 0.283 | 0.817 | 0.451 | 0.780 | HFT | 0.937 | 0.700 | 0.923 | 0.552 | 0.937 | 0.448 | 0.964 | 0.648 | 0.933 | 0.554 | 0.960 | 0.706 | 0.942 | LG | 0.814 | 0.515 | 0.859 | 0.464 | 0.904 | 0.379 | 0.635 | 0.145 | 0.862 | 0.499 | 0.884 | 0.572 | 0.826 | ERDEM | 0.882 | 0.617 | 0.912 | 0.555 | 0.962 | 0.526 | 0.840 | 0.349 | 0.923 | 0.575 | 0.916 | 0.635 | 0.906 | MQDCT | 0.840 | 0.564 | 0.894 | 0.517 | 0.947 | 0.489 | 0.806 | 0.311 | 0.888 | 0.524 | 0.923 | 0.670 | 0.883 | AWS | 0.894 | 0.625 | 0.922 | 0.569 | 0.957 | 0.480 | 0.891 | 0.402 | 0.920 | 0.534 | 0.925 | 0.655 | 0.918 | Proposed | 0.920 | 0.653 | 0.943 | 0.596 | 0.976 | 0.555 | 0.934 | 0.516 | 0.942 | 0.532 | 0.956 | 0.688 | 0.945 |
|
|
The top ranked model is in bold font and the 2nd ranked model is in italic font. The overall best average AUC score is given by our proposed model.
|