Shoot high-quality color images using dual-lens system with monochrome and color cameras

In the dual-lens system with monochrome and color cameras, the gray image captured by the monochrome camera has better quality than the color image from the color camera, but does not have color information. To get high-quality color images, it is desired to colorize the gray image with the color image as reference. Due to occlusions, the colorization will inevitably fail in some cases. Thus, evaluating the colorization quality is also of great importance. We solve both problems in this paper. For colorization, we propose a gray-color correspondence prior, i.e. in local regions, if two patches are similar in the gray channel, it is very often that the two pixels centered at these two patches have similar colors. Based on this prior, a deep learning based and coarse-to-fine colorization method is proposed. For evaluating the colorization quality, we propose a symmetry colorization based evaluation method. Experimental results show that our method could largely outperform the state-of-the-art methods and is also efficient in computation.