Skip to main content

Advertisement

Table 5 Comparison of the SSIM value (0.−) for inpainting of images with various corruption levels based on analytic dictionaries, DCT and MCA, and dictionaries learned on all available corrupted image patches, BPFA, BPFAomp, wKSVD and ITKrMM (modified inpainting is marked with a *)

From: Fast dictionary learning from incomplete data

  Algorithm Bar. Cam. Hou. Man. Pepp. Pir.
30% Corr. Noisy Im. 1689 2367 0831 1604 1630 1619
  DCT64 9822 9638 9813 9373 9859 9691
  DCT144 9812 9629 9814 9413 9858 9698
  MCA 9695 9500 9658 9218 9654 9552
  BPFA 9438 9388 9608 8855 9710 9452
  BPFAomp 9590 9564 9771 9400 9820 9688
  wKSVD 9659 9551 9772 9341 9812 9662
  ITKrMM1 9719 9597 9819 9418 9840 9701
  ITKrMM3 9798 9612 9823 9428 9852 9729
50% Corr. Noisy Im. 1002 1617 0478 0899 1015 0946
  DCT64 9503 9207 9542 8469 9647 9215
  DCT144 9486 9194 9548 8566 9645 9239
  MCA 9424 9173 9484 8408 9511 9157
  BPFA 9284 9141 9508 8187 9612 9171
  BPFAomp 9384 9222 9607 8782 9686 9377
  wKSVD 9439 9257 9607 8786 9683 9363
  ITKrMM1 9514 9281 9651 8816 9688 9374
  ITKrMM3 9588 9281 9657 8825 9695 9392
70% Corr. Noisy Im. 0552 0981 0268 0459 0575 0515
  DCT64 8703 8411 8976 6950 9115 8224
  DCT144 8682 8389 8982 7081 9098 8275
  MCA 8807 8592 9070 7056 9152 8363
  BPFA 8783 8587 9231 7104 9366 8604
  BPFAomp 8828 8600 9235 7614 9374 8713
  BPFAomp* 8834 8602 9238 7615 9387 8740
  wKSVD 8877 8648 9286 7616 9380 8750
  wKSVD* 8885 8676 9290 7615 9392 8757
  ITKrMM1 8937 8661 9306 7582 9222 8720
  ITKrMM1* 9002 8749 9323 7583 9404 8753
  1. The results are averaged over 5 random masks and initialisations. The best result for each setting is marked in bold