An Exemplar-Based Method for Automatic Visual Editing and Retouching of Fine Art Reproduction

摘要

The advance in camera and imaging technologies has made digital archiving and conservation of artworks possible in museums. Visual editing and retouching are usually made by experts in museums to match the reproduction with the original more closely. While effective, visual editing and retouching is time-consuming and labor-intensive. We addressed this problem by learning from the adjustments made by observers, and automatically making visual editing and retouching on incoming paintings. The evaluation results suggest that the image adjusted by our model is significantly better than the average of the images adjusted by observers.