Learning a No-Reference Quality Assessment Model of Enhanced Images With Big Data

Abstract

In this paper, we investigate into the problem of image quality assessment (IQA) and enhancement via machine learning. This issue has long attracted a wide range of attention in computational intelligence and image processing communities, since, for many practical applications, e.g., object detection and recognition, raw images are usually needed to be appropriately enhanced to raise the visual quality (e.g., visibility and contrast). In fact, proper enhancement can noticeably improve the quality of input images, even better than originally captured images, which are generally thought to be of the best quality. In this paper, we present two most important contributions. The first contribution is to develop a new no-reference (NR) IQA model. Given an image, our quality measure first extracts 17 features through analysis of contrast, sharpness, brightness and more, and then yields a measure of visual quality using a regression module, which is learned with big-data training samples that are much bigger than the size of relevant image data sets. The results of experiments on nine data sets validate the superiority and efficiency of our blind metric compared with typical state-of-the-art full-reference, reduced-reference and NA IQA methods. The second contribution is that a robust image enhancement framework is established based on quality optimization. For an input image, by the guidance of the proposed NR-IQA measure, we conduct histogram modification to successively rectify image brightness and contrast to a proper level. Thorough tests demonstrate that our framework can well enhance natural images, low-contrast images, low-light images, and dehazed images. The source code will be released at https://sites.google.com/site/guke198701/publications.

Encoding Visual Sensitivity by MaxPol Convolution Filters for Image Sharpness Assessment

“…Variational maps are also reported as a useful feature. The map has been used in several methods of NR metric development in conjunction with other feature methods [12], [31]. The transfer image domain into total variation and local variation for image analysis is another example to develop NR-ISA metric [4], [10].…”

“…In discrete computational modeling, many existing non-reference (NR) image sharpness assessment (ISA) approaches make use of this fact to grade the sharpness level of images. Examples include but are not limited to the gradient map approaches which approximate HVS response by finite impulse response (FIR) filters [4]-[6], [6]- [9], variational methods [4], [10], [11] and the contrast map techniques [12]- [15]. However, these existing methods are suboptimal in the sense that they cannot fully manifest the reality of the HVS response.…”

A Fast SVM-Based Tongue’s Colour Classification Aided by k-Means Clustering Identifiers and Colour Attributes as Computer-Assisted Tool for Tongue Diagnosis

“…Moreover, to determine the most important features that can exclusively contribute to improve performance of the classifier is very crucial. Some related works are reported in [6–9] using hybrid classifiers to sift out several redundant features and adopted only the most meaningful instances to boost up the performance of the classifier in terms of its response time and classification accuracy. Through this motivation, this technique will reduce the number of training examples and lead to the number of support vectors reduction, thus speeding up the classifier's response time.…”

“…Similar researches in [6–9] have been discussed to sift out redundant information from the samples used in the classifier during data selection. Nevertheless, some of these works did not consider the basis theory of the sample selection and the possibility of misclassification or misclustering during sample selection.…”

Improving the Performance of No-Reference Image Quality Assessment Algorithm for Contrast-Distorted Images Using Natural Scene Statistics

“…In [26], Gu et al investigated the problem of image quality assessment (IQA) and enhancement using machine learning. In their work, they developed a new NR-IQA model by extracting 17 features from the given image by analyzing contrast, sharpness, brightness, among other features.…”