Pairwise Learning to Rank for Image Quality Assessment

Abstract

Because the pairwise comparison is a natural and effective way to obtain subjective image quality scores, we propose an objective full-reference image quality assessment (FR-IQA) index based on pairwise learning to rank (PLR). We first compose a large number of pairs of images, extract their features, and compute their preference labels as training labels. We then obtain a pairwise preference model by training a binary classifier using the features and labels. Because image quality is affected by the masking effect, we propose extracting frequency-aware quality features by adapting state-of-the-art IQA metrics. The learned pairwise preference model is then used to predict the preference between pairs of images in the testing dataset. The quality of each image is computed as the number of preferences. Experimental results on four IQA databases validate that the proposed PLR-based IQA index achieves higher consistency with human subjective evaluation than the state-of-the-art IQA metrics.

Analysis of full reference quality metrics for image transmission over a MIMO OWC channel under varying turbulent conditions

“…This method is easy to evaluate for real time scenario. However, since human vision system is still evolving due to research and understanding of human biology, Objective IQA has its own limitations 28,29 . Objective IQA can further be classified into full reference, no reference and semi‐reference metrics 26,30 .…”

“…The full reference IQA considers undistorted image as ground truth reference and compares it with the output image from system under consideration. Various full reference IQA parameters are listed in previous works 26–30,31 . presented image transmission on mobile wireless sensor networks and evaluated IQA metrics such as MSE, PSNR and WSNR 32 .…”

“…However, since human vision system is still evolving due to research and understanding of human biology, Objective IQA has its own limitations. 28,29 Objective IQA can further be classified into full reference, no reference and semi-reference metrics. 26,30 The full reference IQA considers undistorted image as ground truth reference and compares it with the output image from system under consideration.…”

“…Various full reference IQA parameters are listed in previous works. [26][27][28][29][30]31 presented image transmission on mobile wireless sensor networks and evaluated IQA metrics such as MSE, PSNR and WSNR. 32 utilized SSIM as a metric to model end to end distortion for video transmission.…”

A Pairwise Deep Ranking Model for Relative Assessment of Parkinson’s Disease Patients From Gait Signals

“…In our study, we address a completely different problem; we aim to predict if the first sample is greater than the second sample in terms of an independent continuous label, which measures any quantity of the input signal. This problem has been tackled very recently for pairwise 'ranking' of image data in terms of their quality [22]. We have also recently seen some applications of pairwise ranking in video data for action quality assessment in sport activities as well [23].…”