Study On Quality Assessment Method Of Halftone Images Based On Texture Visual Characteristic

Digital halftoning is a nonlinear imaging process that processes a multi-gray-scale continuous tone image into a binary-valued halftone image through different halftoning algorithms.These various halftone images produced by different halftoning algorithms exhibit different characteristics,and the visual perception of the reproduced images is different.Accurate and effective evaluation of halftone image quality is of important practical significance to the improvement and optimization of the halftone technology.According to whether or not it depends on the original image information,the quality evaluation methods of the halftone images are divided into reference quality assessment and non-reference quality assessment.The quality evaluation of digital halftone images is generally based on reference quality assessment.Commonly used methods for evaluating the quality of halftone images include structural similarity index,block peak to signal-to-noise ratio,weighted signal-to-noise ratio and weighted structural similarity.These halftone image quality evaluation methods either take into account the structural correlation of the image,or take into account the binary nature of the halftone image,the noise characteristics of halftone image,or the human eye's low-pass filtering.However,the evaluation results are often inconsistent with the subjective evaluation results because of the failure to fully consider the factors affecting the quality reproduction of the halftone images.Halftone images contain two types of texture structure distortions:worm-like defects which appear in the highlights and shadows of the image,and stripe-like defects which appear in the middle of the image.These two types of structure distortions seriously affect the copy quality of halftone images.Therefore,three major influencing factors should be considered for the quality evaluation method of halftone images:human eye vision characteristics,the characteristics of halftone image(binary,texture)and image structure similarity.In this paper,we propose a halftone image quality evaluation method based on texture vision characteristics.First,the low pass filtering of human eye vision system is considered.Before assessing the quality of halftone image,continuous-tone image and its halftone image were convoluted with Point Spread Function separately.Furthermore,according to the texture characteristics and the ambiguity of the halftone image,weight values of different gradation inputs are first designed according to the visual range of gradation that they appear,and then the weighted variance of the blocks is used to measure local texture contast.Finally,a weighted summation based on the image information is obtained to obtain a texture distortion quality evaluation value suitable for the halftone image.Matlab_R2012a was used to do the experiments in this paper.The halftone images which are output by different halftoning algorithms are evaluated by the evaluation method proposed in the article.At the same time,the article designed a subjective psychological visual experiment and objective method comparison experiment to evaluate the effectiveness of the proposed quality assessment method.The experimental results show that the quality evaluation method of digital halftone images proposed in this paper is more targeted and effective than the classical objective quality assessment methods,and is consistent with the subjective visual perception.