针对交互式图像色彩编辑对交互采样数据量要求较高的问题, 提出了一种基于分块特征的交互式图像色彩编辑方法。图像色彩的变化是由于原始图像中每一分块的特征的变化所致, 首先, 将图像分块, 以图像块在LUV坐标基下, UV与L存在线性关系且在小的分块内UV与L存在相同的线性关系系数为基础, 将此线性系数作为分块图像的特征建立图像色彩扩散优化模型。其次, 利用图像变换前后内容的结构不变性, 通过LLE简化优化模型, 最后, 利用稀疏线性代数方程组进行求解。实验结果表明, 文中方法在用户低采样的条件下较好地保持了图像结构, 自然渲染了图像色彩。

In order to meet the high demand for sampling data in interactive image color editing, a new interactive image color editing method based on block feature was proposed. The change of image color was caused by the change of features of each block in the original image. Firstly, the image was divided into blocks in the LUV coordinate, and the linear correlation coefficient between UV and L was found in small blocks. The correlation coefficient was used as the feature of the block image to establish the optimal model of image color diffusion. By using the structural similarity before and after image transformation, LLE was used to simplify the optimization model. Finally, the sparse linear algebraic equations were used to solve the problem. The experimental results show that the proposed method can preserve the structure of the image and render the image color naturally in the condition of low sampling.