Learning Data-driven Reflectance Priors for Intrinsic Image Decomposition
This work addresses the problem of separating reflectance and shading in images for computer vision applications, representing an incremental improvement over existing methods.
The authors tackled intrinsic image decomposition by developing a data-driven reflectance prior from human annotations of patch brightness comparisons, which they integrated into existing energy minimization frameworks. Their method outperformed the state-of-the-art approach of Bell et al. on decomposition and relighting tasks, particularly in challenging lighting conditions.
We propose a data-driven approach for intrinsic image decomposition, which is the process of inferring the confounding factors of reflectance and shading in an image. We pose this as a two-stage learning problem. First, we train a model to predict relative reflectance ordering between image patches (`brighter', `darker', `same') from large-scale human annotations, producing a data-driven reflectance prior. Second, we show how to naturally integrate this learned prior into existing energy minimization frameworks for intrinsic image decomposition. We compare our method to the state-of-the-art approach of Bell et al. on both decomposition and image relighting tasks, demonstrating the benefits of the simple relative reflectance prior, especially for scenes under challenging lighting conditions.