CAJ | 학술논문

为提高多描述编码的性能，研究了一种基于分级量化压缩感知（compressive sensing，CS）的多描述编码方法。通过交织抽取把图像分块得到几个子图像，对每个子图像进行分块压缩感知形成多个描述码流，接收端通过求解优化问题重建原图像。为了提高比特利用率，使用分级量化方法对测量值进行量化。实验结果表明，在保证图像恢复质量的前提下，该方法可以节省编码所需比特数，不影响多描述编码的抗丢包性能。
위제고다묘술편마적성능，연구료일충기우분급양화압축감지（compressive sensing，CS）적다묘술편마방법。통과교직추취파도상분괴득도궤개자도상，대매개자도상진행분괴압축감지형성다개묘술마류，접수단통과구해우화문제중건원도상。위료제고비특이용솔，사용분급양화방법대측량치진행양화。실험결과표명，재보증도상회복질량적전제하，해방법가이절성편마소수비특수，불영향다묘술편마적항주포성능。
To improve the performance of multiple description coding,a multiple description coding method based on scalar quan-tized compressed sensing was investigated.This method first partitioned an image into several sub-images using interleaving ex-traction,then measured each sub-image with block compressive sensing and formed multiple descriptions.At the decoding termi-nal,it reconstructed the original image by solving an optimization problem.Scalar quantization (SQ)was used to quantify the measurements to enlarge the rate of bits utilization.Experimental results showed that,on the premise of ensuring image restora-tion quality,the SQ method could save the number of bits required by encoding without affecting the ability against packet loss.