No-reference tone mapping image quality evaluation algorithm based on clustering perception

Abstract

The invention discloses a no-reference tone mapping image quality evaluation algorithm based on clustering perception, which comprises the following steps of: A, extracting clustering perception characteristics, namely respectively extracting two characteristics of area ratio and information entropy; B, extracting salient region features, namely extracting two features of block proportion and information entropy; C, extracting naturalness features, wherein the naturalness feature extraction comprises brightness statistical features and color channel statistical features; and step D, performingregression on all the features by using a machine learning method, so that an image quality evaluation score can be obtained. Compared with the prior art, the algorithm has the beneficial effects that a reference image is not needed when the test image is evaluated; a partitioning result of K-means clustering is improved, so that the precision of the algorithm is improved; the matrix non-negativedecomposition speed is high, and the detection method has a very good operation speed; the naturalness characteristics are combined with the brightness naturalness and the color naturalness, and theSROCC performance of the algorithm is greatly improved compared with single naturalness.

Description

technical field [0001] The invention relates to an image quality evaluation algorithm, in particular to a no-reference tone mapping image quality evaluation algorithm based on cluster perception. Background technique [0002] A high dynamic range (High Dynamic Range, HDR) image can represent a larger brightness range than a low dynamic range (Low Dynamic Range, LDR) image, and its brightness range is about 10 -4 cd / m 2 ~10 5 cd / m 2 . The dynamic range that the LDR image can represent does not exceed 3 orders of magnitude, but the dynamic range that the human visual system can accept in real scenes can reach 6 orders of magnitude. Therefore, HDR images feel more realistic and attractive to users. With the development of imaging and computer graphics technology, HDR images are getting easier and easier to obtain. However, HDR display devices are relatively expensive and out of the reach of ordinary consumers. To solve this problem, many Tone-mapped Operators (TMO) have ...

AI Technical Summary

Problems solved by technology

Since TMI distortion usually does not appear blurring, block effect and other types of distortion, traditional quality evaluation algorithms are not suitable for evaluating TMI

The quality evaluation method of ordinary (low dynamic) image is not suitable for tone mapping image quality evaluation, because the distortion type of tone mapping image is different from that of low dynamic image, and the distortion type of low dynamic image is mainly block effect, blurring, ringing effect, while the distortion type of the tone-map

Images