BokehMe: When Neural Rendering Meets Classical Rendering

TL;DR

BokehMe is a hybrid rendering framework combining neural and classical methods to produce high-quality, adjustable bokeh effects from a single image and disparity map, effectively correcting errors and handling various blur sizes.

Contribution

It introduces a novel hybrid approach that integrates a classical scattering-based renderer with a neural network to improve bokeh rendering quality and robustness.

Findings

Outperforms previous methods on synthetic and real data

Effectively handles imperfect disparity maps

Provides adjustable bokeh effects with high realism

Abstract

We propose BokehMe, a hybrid bokeh rendering framework that marries a neural renderer with a classical physically motivated renderer. Given a single image and a potentially imperfect disparity map, BokehMe generates high-resolution photo-realistic bokeh effects with adjustable blur size, focal plane, and aperture shape. To this end, we analyze the errors from the classical scattering-based method and derive a formulation to calculate an error map. Based on this formulation, we implement the classical renderer by a scattering-based method and propose a two-stage neural renderer to fix the erroneous areas from the classical renderer. The neural renderer employs a dynamic multi-scale scheme to efficiently handle arbitrary blur sizes, and it is trained to handle imperfect disparity input. Experiments show that our method compares favorably against previous methods on both synthetic image data and real image data with predicted disparity. A user study is further conducted to validate the advantage of our method.