Skip to main content
. 2017 Jan 18;17(1):174. doi: 10.3390/s17010174

Figure 5.

Figure 5

Comparison of results on one test image from [4]. From top-left to bottom-right: (a) latent sharp image, deblurred versions with (b) the truth blur kernel, estimated blur kernels generated by (c) Fergus et al. [7], (d) Xu and Jia [11], (e) Cho and Lee [12], (f) Levin et al. [58], (g) Pan and Su [13] and (h) our proposed method, respectively. The estimated blur kernels are illustrated in the upper-left panels.