视觉显著性导向的图像压缩感知测量与重建

Visual saliency oriented compressive sensing measurement and reconstruction of images

提出利用视觉显著性指导图像压缩感知的自适应测量与重建的算法.考虑到感知端不可负载过多的计算量,采用亮度对比度计算输入全采样图像的显著度,并根据块显著度实现自适应测量;重建端利用动态变化的块测量率重新估算块显著度,并以此加权重建模型的目标函数,集中优化高显著块.实验结果表明:与传统算法相比,所提算法重建图像的整体客观质量更优,且可更好地保护边缘与纹理等重要细节,主观视觉质量良好,同时保证了较低的测量与重建计算复杂度.

Luminance contrast was used to compute the saliency value of input all-sampling image,and the adaptive measurement wais realized depending on the saliency value of image block.At the reconstruction side,the saliency value of each block was estimated by using varying block measurement rates,and then these saliency values were used to weight the objective function of reconstruction model in order to enforce the quality improvements of high-saliency blocks.Experimental results indicate that the reconstructed image by the proposed algorithm has a better objective quality when comparing with several traditional ones,and its edge and texture details are better preserved,which guarantees the better subjective visual quality.Besides,the proposed method has a low computational complexity of measurement and reconstruction.