Supervised Learning of Digital image restoration based on Quantization Nearest Neighbor algorithm

TL;DR

This paper introduces a supervised image restoration algorithm that leverages learned priors and quantization-based nearest neighbor techniques to effectively restore degraded images by estimating high-frequency details.

Contribution

It presents a novel supervised approach combining quantization and nearest neighbor algorithms for image restoration, differing from traditional methods by learning priors from similar images.

Findings

Effective restoration of degraded images with high accuracy.

Ability to identify and adapt to different blurring functions.

Improved computational efficiency through quantization techniques.

Abstract

In this paper, an algorithm is proposed for Image Restoration. Such algorithm is different from the traditional approaches in this area, by utilizing priors that are learned from similar images. Original images and their degraded versions by the known degradation operators are utilized for designing the Quantization. The code vectors are designed using the blurred images. For each such vector, the high frequency information obtained from the original images is also available. During restoration, the high frequency information of a given degraded image is estimated from its low frequency information based on the artificial noise. For the restoration problem, a number of techniques are designed corresponding to various versions of the blurring function. Given a noisy and blurred image, one of the techniques is chosen based on a similarity measure, therefore providing the identification of the blur. To make the restoration process computationally efficient, the Quantization Nearest Neighborhood approaches are utilized.