Abstract

Photographs taken through glass windows often contain both the desired scene and undesired reflflections. Separating the reflflection and transmission layers is an important but ill-posed problem that has both aesthetic and practical applications. In this work, we introduce the use of ghosting cues that exploit asymmetry between the layers, thereby helping to reduce the ill-posedness of the problem. These cues arise from shifted double reflflections of the reflflected scene off the glass surface. In double-pane windows, each pane reflflects shifted and attenuated versions of objects on the same side of the glass as the camera. For single-pane windows, ghosting cues arise from shifted reflflections on the two surfaces of the glass pane. Even though the ghosting is sometimes barely perceptible by humans, we can still exploit the cue for layer separation. In this work, we model the ghosted reflflection using a double-impulse convolution kernel, and automatically estimate the spatial separation and relative attenuation of the ghosted reflflection components. To separate the layers, we propose an algorithm that uses a Gaussian Mixture Model for regularization. Our method is automatic and requires only a single input image. We demonstrate that our approach removes a large fraction of reflflections on both synthetic and real-world inputs.