Reflection Separation via Multi-bounce Polarization State Tracing

Abstract

Reflection removal from photographs is an important task in computational photography, but also for computer vision tasks that involve imaging through windows and similar settings. Traditionally, the problem is approached as a single reflection removal problem under very controlled scenarios. In this paper we aim to generalize the reflection removal to real-world scenarios with more complicated light interactions. To this end, we propose a simple yet efficient learning framework for supervised image reflection separation with a polarization-guided raytracing model and loss function design. Instead of a conventional image sensor, we use a polarization sensor that instantaneously captures four linearly polarized photos of the scene in the same image. Through a combination of a new polarization-guided image formation model and a novel supervised learning framework for the interpretation of a raytracing image formation model, a general method is obtained to tackle general image reflection removal problems. We demonstrate our method with extensive experiments on both real and synthetic data and demonstrate the unprecedented quality of image reconstructions.

Semantic Guided Single Image Reflection Removal

“…Punnappurath et al [40] use dual pixel sensors to capture two sub-aperture views of the scene, and then ind defocus disparity clues for relection removal. Many works [20,22,25,56,57] explore the polarization and take multiple images to solve the optimal separation through angle ilter. Lei et al [21] propose to use a pair of lash and no-lash images as input.…”

Single image reflection removal with camera calibration

“…Lei [75] utilize the polarization information and proposes a polarized reflection removal model with a two-stage architecture to solve the ill-posed problem. Li [76] further proposes to a polarization-guided raytracing model and loss function design to separate reflection images.…”