An Optical physics inspired CNN approach for intrinsic image decomposition
This addresses the lack of unsupervised learning approaches for intrinsic image decomposition, which is important for computer vision applications, though it appears incremental as it builds on existing deep learning methods.
The authors tackled the problem of intrinsic image decomposition by proposing a neural network architecture that uses physics-based parameters to separate reflectance and shading from a single image without ground truth, showing it outperforms existing deep learning techniques with significant improvements in efficacy.
Intrinsic Image Decomposition is an open problem of generating the constituents of an image. Generating reflectance and shading from a single image is a challenging task specifically when there is no ground truth. There is a lack of unsupervised learning approaches for decomposing an image into reflectance and shading using a single image. We propose a neural network architecture capable of this decomposition using physics-based parameters derived from the image. Through experimental results, we show that (a) the proposed methodology outperforms the existing deep learning-based IID techniques and (b) the derived parameters improve the efficacy significantly. We conclude with a closer analysis of the results (numerical and example images) showing several avenues for improvement.