Class-adapted blind image deblurring (by Marina)

In the previous post, I introduced the concept of blind image deblurring: the problem when we have an observed blurred image and we want to find an underlying sharp image and a blurring filter.  Also, I mentioned that there are different types of blurring filters and they have different characteristics. What I did not mention is that there are also different types of sharp images (figure 1).

Figure 1: Different types of images.

Different types of sharp images can have completely different structure. For example, look at a (document like) text image which has sharp transitions and usually comes in two tones, black and white. On the other side, a natural image has smooth regions (like sky or lake), texture (bridge, trees) and some sharp edges (boat). An image of a face can have small texture (like area below the eyes), that is very difficult to estimate ones when is blurred.

As we have a problem where we need to estimate two unknowns (an underlying sharp image and a blurring filter) from only one observation (a blurred image), we usually introduce some kind of prior knowledge on both the sharp image and the blurring filter. A prior knowledge of image can represent a characteristic or a set of characteristics about sharp images itself. The similar is for blurring filters. But, as you know so far, there are different types of images and different types of blurs, so how to find a unique prior knowledge (one for an image and one for a blur) that can cover these differences?

We proposed one solution in the paper presented at the International Conference on Image Processing (ICIP) 2017. The main idea behind the proposal is that we can learn an image prior for different types of images. For example, if we know that we have a blurred image of text, we can use sharp images that contain text (other images) to learn some underlying characteristics of text images.

Figure 2: Clean image of text, blurred image of text and our estimated image.

More information can be found here: https://arxiv.org/abs/1709.01710

Image deblurring (by Marina)

First, I want to apologize for our (little bit longer) disappearance from the blog. I hope that with next several posts you will see the reason behind it.
Here, I want to start with an introduction of my specific topic, blind image deblurring. After the introduction, following posts will be about some of my recent publications, a little bit more technical, but still, I hope, interesting.

What is image deblurring?
It is a process where you have an observed blurred image and you want to estimate, to find, a sharp original image hidden “behind” the blurred image (figure 1).

Figure 1: Image deblurring.

The blurring of the image can have different reasons, like camera shake or out-of-focus, and it is usually described with the point spread function (PSF) or a blurring operator (filter). As there are different reasons for blurry image, there are different blurring filters (figure 2).

Figure 2: Blurring filters: Gaussian, linear motion, out-of-focus, uniform and nonlinear motion blur.

There are two types of image deblurring:
1.    Non-blind image deblurring – where we know how the blurring filter looks like.
2.    Blind image deblurring – when we have a partial knowledge or no knowledge at all about the blurring filter.

The second problem is more realistic one (because usually, we do not know how the blurring occurs) and it is the one that I am interested in. Also, blind image deblurring is more complicated to solve, because in this case we only have a blurred image (one that we observed), and we need to find an underlying sharp image and a blurring filter at the same time. Also, an additional problem that we have is a presence of noise, like we explained in one of the previous posts, that makes our job even harder.