图像分块压缩感知中的自适应测量率设定方法

发布时间：2018-07-11 11:31

本文选题：分块压缩感知 + 压缩成像　；参考：《通信学报》2014年07期

【摘要】：传统的图像分块压缩感知(BCS,block compressed sensing)以相同的测量率对各块进行测量,但由于图像的空间特性不同,在重构图像时出现了块效应。通过自适应为各块设定不同的测量率,该问题可得到有效的解决。然而,已有的自适应测量率设定法需要在采集端获得原始数字图像,这在实际的压缩成像(CI,compressive imaging)设备中无法实现。为了克服这一缺陷,提出了一种更易于通过硬件实现的自适应测量率设定法。该方法利用在采集端可获得的CS测量值直接在测量域中估计各图像块的样本方差,再根据各块样本方差自适应地为每块设定测量率并实现码率控制。仿真实验结果表明,该方案重构图像的质量优于非自适应方案,但由于测量域估计块样本方差存在偏差,使其与直接利用块样本方差真实值的自适应方案相比,仍具有一定差距。
[Abstract]:The traditional block compression sensing (BCS) block compressed sensing) measures each block at the same measurement rate, but because of the different spatial characteristics of the image, there is a block effect in the reconstruction of the image. By adaptively setting different measurement rates for each block, the problem can be solved effectively. However, the existing adaptive measurement rate setting method needs to obtain the original digital image at the acquisition end, which can not be realized in the actual compression imaging (imaging) equipment. In order to overcome this defect, an adaptive rate setting method is proposed, which is more easily realized by hardware. In this method, the sample variance of each image block is estimated directly in the measurement domain by using the CS measurements available at the acquisition end, and then the measurement rate is set adaptively for each block according to the sample variance of each block and the bit rate control is realized. The simulation results show that the quality of reconstructed image is better than that of the non-adaptive scheme. However, because of the deviation in the estimation of the block sample variance in the measurement domain, it is compared with the adaptive scheme which directly uses the real value of the block sample variance. There is still a certain gap.
【作者单位】：南京邮电大学江苏省图像处理与图像通信重点实验室;
【基金】：国家自然科学基金资助项目(61071091,61271240) 江苏省研究生创新计划基金资助项目(CXZZ12_0466,CXZZ11_0390) 江苏省高校自然科学研究基金资助项目(12KJB510019) 南京邮电大学校科研基金资助项目(NY212015) 湖北省教育厅科研重点基金资助项目(D20121408)~~
【分类号】：TN911.73

【相似文献】