Exposure Fusion

A Python module for fast HDR image processing: it aligns a bracketed set of exposures and blends them into one well-exposed image.

2022Open source, MIT

Code Original paper

**Exposure Fusion.** An exposure-fused photograph of a desk at night with a laptop, a statuette, and a lit lamp.

An implementation of the exposure fusion algorithm, which merges photographs taken at different exposures into a single image without building an HDR radiance map first. It adds an alignment step to the original paper's method, which improves the result when the photographs in a sequence are offset from each other.

The repository has the ExposureFusion class, which works like a PyTorch module (create it once, then call it on a list of images), and a small Tkinter interface built on it. It started as a course project for Computer Graphics at Sapienza University of Rome.