基于背景消除的小波域视频质量评价模型

Video Quality Assessment Model in Wavelet Domain Based on Background Subtraction

为降低视频质量评价的复杂度并提高准确度,提出一种基于背景消除的小波域视频质量评价模型(BSWQ).首先,利用小波分解计算边缘系数相似性来获取全局质量;其次,以背景消除法提取视频帧组的失真和运动信息来构建时空域视频块,通过对视频块的边缘系数相似性加权获取每个帧组的质量,然后利用最差质量汇聚策略获取视频序列的局部质量;最后,将全局质量和局部质量求均值开方的倒数得到总体的客观质量.分别在LIVE和CSIQ视频库上对所提算法进行测试,与经典的评价方法 MOVIE相比,BSWQ的SROCC值平均提高了4.2%,,并且复杂度是MOVIE的0.77%,.

In order to reduce the computational complexity and improve the accuracy of video quality assessment,a video quality assessment model in wavelet domain based on background subtraction（BSWQ）is proposed.Firstly,the discrete wavelet transform was applied to calculate the edge similarity to obtain the global quality.Then,spatiotemporal domain video blocks were constructed by using the motion and distortion information extracted from background subtraction.The quality of each group of frames was obtained by weighting the edge similarity indexes of video block,and the worst-case pooling strategy was applied to obtain the local video quality of video sequence.Finally,the total quality was measured by the reciprocal of the root mean of the global quality and local quality.The proposed metric is tested respectively on the LIVE and CSIQ video databases.Compared with classical evaluation method MOVIE,the SROCC of BSWQ is increased by an average of 4.2%,,and the complexity of BSWQ is much lower,only about 0.77%, of that of MOVIE.