Dynamic range reduction using Laplacian pyramids
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
src
.gitignore
LICENSE
Makefile
README.md

README.md

FloatingPyramids

FloatingPyramids performs dynamic range reduction on images using Laplacian pyramids.

The algorithms used are described in the papers Local Laplacian Filters: Edge-aware Image Processing with a Laplacian Pyramid and Fast and Robust Pyramid-based Image Processing.

Dependencies

FloatingPyramids is written in C++ and CUDA and requires CUDA 5.0 or later, OpenCV, and a Linux-compatible build environment.

Installation

To compile FloatingPyramids, edit the Makefile and adjust the variables for your environment (most importantly, choose CUDA_ARCH to match the compute capability of your GPU). Then run make; the resulting executable will be named bin/pyramids.

Usage

pyramids <input_image> <output_prefix>

Note that <input_image> currently must have dimensions that are powers of 2. The monochrome output is written to the file <output_prefix>_final.png, and an experimental color output is written to colorout.png.