Example-based color transfer with Gaussian mixture modeling

Color transfer, which plays a key role in image editing, has attracted noticeable attention recently. It remained challenge to date due various issues such as time-consuming manual adjustments and prior segmentation issues. In this paper, we propose model color transfer under probability framework cast it parameter estimation problem. particular, relate the transferred with example Gaussian Mixture Model (GMM) regard GMM centroids. We employ Expectation-Maximization (EM) algorithm (E-step M-step) for optimization. To better preserve gradient information, introduce Laplacian based regularization term objective function at M-step is solved by deriving descent algorithm. Given input of source an image, our method able generate multiple results increasing EM iterations. Extensive experiments show that approach generally outperforms other competitive methods, both visually quantitatively.