Image Enhancement Using Vector Quantisation Based Interpolation

Abstract

We present a novel method of image expansion using vector quantisation. The algorithm is inspired by fractal coding and uses a statistical model of the relationship between details at different scales of the image to interpolate detail at one octave above the highest spatial frequency in the original image. Our method aims at overcoming the drawbacks associated with traditional approaches such as pixel interpolation, which smoothes the scaled-up images, or fractal coding, which bears high computational cost and has limited use due to patent restrictions. The proposed method is able to regenerate plausible image detail that was irretrievable when traditional approaches are used. The vector quantisation-based method outperforms conventional approaches in terms of both objective and subjective evaluations.

A comprehensive assessment of the structural similarity index

“…The product of several years of research [29,30,32,33,38] itself, this metric has been formulated and revised a number of times. It has been employed in image restoration [3], video quality monitoring [4,6,35,36], image enhancement [5], video compression [15,25], visual cognition [18], and imaging coding [34]. Further applications are described in [31].…”

Image Enhancement Methods: A Review

“…Existing interpolation techniques include adaptive Sub-Pixel [20], high-frequency sub-band by discrete wavelet [21], bi-linear/bi-cubic [20], Vector Quantisation [22] and dual tree-complex wavelet [23]. In most cases, the image is enlarged to a scale factor derived from the mean, median or maxima of its neighbouring pixels.…”

Correction of a Noisy Image by a Polynomial Approach and Choice of the Best Image by One of the Polynomial’s Roots.

“…In the field of improving the quality of an image disturbed by noise, many authors have developed classical improvement methods such as: interpolation method which uses various shapes as described by authors [1][2][3][4], histogram equalization method which gives good results in improving color images [5][6][7][8] and medical images [9][10][11]; contrast stretching method which focuses on improving the contrast of an image by "stretching" its range of intensity values to cover a range desired or authorized, presented by the authors [12][13]; compression of the dynamic range [14], partial differential equations method widely used in image filtering and restoration [15][16]; cellular neural networks method known for their success in improving and analyzing medical images [17]; the directional wavelet transformation mainly used for feature extraction, enhancement, denoising, classification and compression [18]. Some methods provide algorithms that apply to areas of an image or its parts all that possess noise.…”