Skip to main content
. 2023 Aug 16;18(12):2349–2356. doi: 10.1007/s11548-023-02974-3

Table 1.

Mosaicking performance in terms of mean ± standard deviation of SSIM5 for all the videos in the dataset between Bano et al. (EM1), SIFT (EM2), ORB (EM3) and the proposed framework without occlusion recovery (EM4)

Video # EM1 EM2 EM3 EM4
1 0.8556±0.0809 0.8241±0.1512 0.8098±0.1170 0.8257±0.0834
2 0.8714±0.0666 0.8441±0.0563 0.6353±0.3585 0.8497±0.0553
3 0.8670±0.1079 0.7658±0.3060 0.5713±0.4197 0.8946±0.0440
4 0.7457±0.1921 0.4526±0.4134 0.7526±0.1773
5 0.9445±0.8720 0.5415±0.4021 0.4911±0.4339 0.9406±0.0316
6 0.8720±0.0646 0.0309±0.1358 0.8750±0.0610
7 0.8133±0.2429 0.8974±0.0333 0.8999±0.0295 0.8455±0.1200
8 0.7603±0.2623 0.8740±0.0710 0.7961±0.2560 0.8463±0.0808
9 0.7947±0.1506 0.5777±0.4129 0.9195±0.0598
10 0.8155±0.1649 0.8209±0.2749 0.9239±0.0653

The highest average values are shown in bold. The values in italic represent mosaics that were discarded after visual inspection due to errors in reconstruction