基于局部对比度保持的色阶重建算法

Tone Reproduction Algorithm Based on Local Contrast Preservation

为了解决现有色阶重建算法存在的明暗对比欠佳以及细节弱化问题,提出一种基于局部对比度保持的色阶重建算法.首先利用二阶范数构建聚类准则,通过动态规划对聚类准则进行最优化,实现全局色阶重建;然后通过局部边缘保持滤波对局部对比度进行估计,根据局部对比度一致性构建线性方程组,方程组的解是保证局部细节特征的重建图像.在TMQI数据库上的实验结果表明,所提算法对应的主观打分均值达到0.898,色阶映射图像质量测度达到0.896,局部均值韦伯对比度达到1.645,均高于现有的色阶重建算法;该算法可以有效地保证重建图像整体明暗效果,重建图像的局部细节较为丰富.

In order to address suboptimal contrast and attenuated detail which are introduced by existing tone reproduction algorithms,a tone reproduction algorithm is proposed based on local contrast preservation.Firstly,the proposed algorithm takes the second-order norm as the clustering criterion.Dynamic programming is employed to find the optimal solutions for global tone reproduction.Secondly,linear equations are deduced by equivalent local contrast that is estimated by local edge-preserving filtering.The solution of the equations is the reproduction image that preserves local details.Comparison experiment results on TMQI database indicate that the mean opinion score,tone-mapped image quality index,and mean local weber contrast of tone mapped image reproduced by the proposed algorithm achieve 0.898,0.896,and 1.645.Thus,the proposed algorithm ensures that the overall contrast of the reproduced image and the reproduced image is full of rich detail on all parts of the picture.