CAJ | 학술논문

为解决数字图像处理中的质量评价问题,从自然图像统计和空域变换角度出发,提出一种无参质量评价方法。首先采用空域变换并结合梯度关系获得图像的统计特性分布；然后采用非对称广义高斯分布进行模拟,在充分考虑非对称因素的情况下求得特征参数,以反映分布特征；最后利用失真图像与原始图像统计分布的差异,将特征参数与参考标准直接进行 K-L 距计算获得评价值。实验结果证明,文中方法无需参考图像,适用于任意失真类型图像的质量评价,与同类方法相比,评价结果与主观DMOS值更具一致性；同时,该方法计算复杂度较低,对Live库中图像运算耗时少于130 ms,具有广泛的实际应用价值。
위해결수자도상처리중적질량평개문제,종자연도상통계화공역변환각도출발,제출일충무삼질량평개방법。수선채용공역변환병결합제도관계획득도상적통계특성분포；연후채용비대칭엄의고사분포진행모의,재충분고필비대칭인소적정황하구득특정삼수,이반영분포특정；최후이용실진도상여원시도상통계분포적차이,장특정삼수여삼고표준직접진행 K-L 거계산획득평개치。실험결과증명,문중방법무수삼고도상,괄용우임의실진류형도상적질량평개,여동류방법상비,평개결과여주관DMOS치경구일치성；동시,해방법계산복잡도교저,대Live고중도상운산모시소우130 ms,구유엄범적실제응용개치。
To solve the problem of image quality assessment, a novel no-reference image quality assessment method is proposed based on natural scene statistics and spatial transformation. Firstly, statistical distribution of image by spatial transformation and grads are got. Secondly, asymmetric generalized Gaussian distribution model is used to simulate the distribution. And by reflection on unsymmetrical character, parameters are calculated to show distribution feature. Finally, diversity between distortion image and original image is compared and assessment value by K-L divergence is computed. Experiments reveal that this method can evaluate pictures with any kind of distortion without reference. The assessment results are better than other methods in accordance with subjective DMOS. At the same time, the complexity of calculation of this method is comparatively lower, the time for simple processing each of images in Live database is less than 130ms, which makes it great and wide value for engineering application.