Dr.Bokeh: DiffeRentiable Occlusion-Aware Bokeh Rendering | IEEE Conference Publication | IEEE Xplore

Dr.Bokeh: DiffeRentiable Occlusion-Aware Bokeh Rendering


Abstract:

Bokeh is widely used in photography to draw attention to the subject while effectively isolating distractions in the background. Computational methods can simulate bokeh ...Show More

Abstract:

Bokeh is widely used in photography to draw attention to the subject while effectively isolating distractions in the background. Computational methods can simulate bokeh effects without relying on a physical camera lens, but the inaccurate lens modeling in existing filtering-based meth-ods leads to artifacts that need post-processing or learning-based methods to fix. We propose Dr.Bokeh, a novel ren-dering method that addresses the issue by directly correcting the defect that violates physics in the current filtering-based bokeh rendering equation. Dr.Bokeh first preprocesses the input RGBD to obtain a layered scene representation. Dr.Bokeh then takes the layered representation and user-defined lens parameters to render photo-realistic lens blur based on the novel occlusion-aware bokeh rendering method. Experiments show that the non-learning based renderer Dr.Bokeh outperforms state-of-the-art bokeh ren-dering algorithms in terms of photo-realism. In addition, extensive quantitative and qualitative evaluations show that the more accurate lens model pushes the limit of depth-from-defocus.
Date of Conference: 16-22 June 2024
Date Added to IEEE Xplore: 16 September 2024
ISBN Information:

ISSN Information:

Conference Location: Seattle, WA, USA

1. Introduction

Bokeh is a physical effect produced by a camera lens system. It refers to the shape and quality of out-of-focus areas in an image. Bokeh brings focus to the in-focus subject and enhances the overall aesthetic quality of the image.

Being occlusion-aware, Dr.Bokeh renders realistic bokeh effects from the bokeh rendering process without post-processing. Compared with the scattering/gathering-based method SteReFo and learning-based method BokehMe, Dr.Bokeh renders natural partial occlusion (red parts). MPIB learns to render a partial occlusion effect but breaks on unseen data (blue parts). Dr.Bokeh is more robust than learning-based methods given the same inputs because the rendering process is physically grounded. Best viewed by zooming in.

Artifacts by inaccurate lens model: color bleeding and partial occlusion are two main artifacts introduced by current inaccurate lens model. Color bleeding means the pixels in the out-of-focus scatter to in-focus regions. Partial occlusion is a semi-transparent effect on the out-of-focus boundary regions, where part of the backgrounds are visible in the background in-focus case. Best viewed by zoom-in.

Contact IEEE to Subscribe

References

References is not available for this document.