Description
Image priors have become a popular tool for image restoration tasks. Good priors have been applied to different tasks such as image denoising , image inpainting and more, yielding excellent results. However, learning good priors from natural images is a daunting task – the high dimensionality of images makes learning, inference and optimization with such priors prohibitively hard. As a result, in many works priors are learned over small image patches. This has the advantage of making computational tasks such as learning, inference and likelihood estimation much easier than working with whole images directly.
We now compare the performance of our method (EPLL+GMM) to image specific methods – which learn from the noisy image itself. We compare to KSVD, BM3D and LLSC which are currently the state-of-the-art in image denoising. The summary of results may be seen in Table 2b. As can be seen, our method is highly competitive with these state-of-the-art method, even though it is generic. Some examples of the results may be seen in Figure 1.
![Examples of denoising using EPLL-GMM compared with state-of-the-art denoising methods - KSVD [3] and LLSC [8]. Note how detail is much better preserved in our method when compared to KSVD. Also note the similarity in performance with our method when compared to LLSC, even though LLSC learn from the noisy image. See supplementary material for more examples.](https://matlab1.com/wp-content/uploads/2018/05/fig7.jpg)
Examples of denoising using EPLL-GMM compared with state-of-the-art denoising methods – KSVD [3] and LLSC [8]. Note how detail is much better preserved in our method when compared
to KSVD. Also note the similarity in performance with our method when compared to LLSC, even though LLSC learn from the noisy image. See supplementary material for more examples.
Estimation and Preservation for Texture Enhanced Image Denoising
Reviews
There are no reviews yet.