A Feature-Enriched Completely Blind Image Quality Evaluator

Abstract

Abstract-Existing blind image quality assessment (BIQA) methods are mostly opinion-aware. They learn regression models from training images with associated human subjective scores to predict the perceptual quality of test images. Such opinion-aware methods, however, require a large amount of training samples with associated human subjective scores and of a variety of distortion types. The BIQA models learned by opinion-aware methods often have weak generalization capability, hereby limiting their usability in practice. By comparison, opinion-unaware methods do not need human subjective scores for training, and thus have greater potential for good generalization capability. Unfortunately, thus far no opinion-unaware BIQA method has shown consistently better quality prediction accuracy than the opinion-aware methods. Here, we aim to develop an opinionunaware BIQA method that can compete with, and perhaps outperform, the existing opinion-aware methods. By integrating the features of natural image statistics derived from multiple cues, we learn a multivariate Gaussian model of image patches from a collection of pristine natural images. Using the learned multivariate Gaussian model, a Bhattacharyya-like distance is used to measure the quality of each image patch, and then an overall quality score is obtained by average pooling. The proposed BIQA method does not need any distorted sample images nor subjective quality scores for training, yet extensive experiments demonstrate its superior quality-prediction performance to the state-of-the-art opinion-aware BIQA methods. The MATLAB source code of our algorithm is publicly available at www.comp.polyu.edu.hk/∼cslzhang/IQA/ILNIQE/ILNIQE.htm.Index Terms-Blind image quality assessment, natural image statistics, multivariate Gaussian.

Deep Activation Pooling for Blind Image Quality Assessment

“…In Section 3.6, we study the influence of each kind of features on the final BIQA. Note that the results of blind/referenceless image spatial quality evaluator (BRISQUE), blind image integrity notator using DCT statistics (BLIINDS2), codebook representation for no-reference image assessment (CORNIA), natural image quality evaluator (NIQE) and integrated local natural image quality evaluator (IL-NIQE) are implemented by [19].…”

“…(a) We utilize a set of high-quality pristine images to train a PMVG model [19]. We first choose the CSM in the 4th convolutional layer of the VGG-19 network to select K high-contrast patches from the 90 pristine images using the high-contrast selection method mentioned in Section 2.2.…”

“…For fair comparison, we extract the same features as in [19] for each high-contrast patch. These features include the MSCN, MSCN products, gradients, log-Gabor filter responses and color-based features.…”

“…Xue et al [18] proposed the quality-aware clustering scheme to assess the quality levels of image patches. Zhang et al [19] proposed the integrated local natural image quality evaluator (IL-NIQE) to predict image quality for opinion-unaware BIQA tasks. Firstly, they utilize five types of features to obtain the feature vectors of pristine naturalistic images, that is, the mean-subtracted contrast normalized (MSCN)-based features, MSCN product-based features, gradient-based features, log-Gabor response-based features and color-based features.…”

Flexible Actuator With Variable Stiffness and Its Decoupling Control Algorithm: Principle Prototype Design and Experimental Verification

“…With regard to 2D algorithms, apply them on two views and each frame, then take the mean value as the approximate value of VQA. In the experimental part of the paper, we randomly select 80% videos for training and 20% for test, and there is no overlap between the two sets [53]. The run time is set to 1200, and we take median value as the final results.…”