Dsrn: An Efficient Deep Network For Image Relighting

Custom and natural lighting conditions can be emulated in images of the scene during post-editing. Extraordinary capabilities of the deep learning framework can be utilized for such purpose. Deep image relighting allows automatic photo enhancement by illumination-specific retouching. Most of the state-of-the-art methods for relighting are run-time intensive and memory inefficient. In this paper, we propose an efficient, real-time framework Deep Stacked Relighting Network (DSRN) for image relighting by utilizing the aggregated features from input image at different scales. Our model is very lightweight with total size of about $42$ MB and has an average inference time of about 0.0116s for image of resolution 1024 x 1024 which is faster as compared to other multi-scale models. Proposed method is quite robust for translating image color-temperature from input image to target image. The method also performs moderately for light gradient generation with respect to the target image. Additionally, we demonstrate that the results further improve when images illuminated from opposite directions are utilized as input.