Fast and easy focus stacking

This project implements a cross-platform tool for focus stacking images. The application takes a set of images captured at different focus distances and combines them so that the complete subject is in focus.

Installation

Binary packages for Windows 10 and Ubuntu 18.04 are available on the GitHub releases tab. Ubuntu packages are also available from this PPA.

Basic usage

In most cases just passing all the images is enough:

 build/focus-stack .../path/to/input/*.jpg

For advanced usage, see --help for list of all options:

Usage: build/focus-stack [options] file1.jpg file2.jpg ...
Options:
  --output=output.jpg           Set output filename
  --reference=0                 Set index of image used as alignment reference (default middle one)
  --global-align                Align directly against reference (default with neighbour image)
  --full-resolution-align       Use full resolution images in alignment (default max 2048 px)
  --no-whitebalance             Don't attempt to correct white balance differences
  --no-contrast                 Don't attempt to correct contrast and exposure differences
  --threads=2                   Select number of threads to use (default number of CPUs + 1)
  --no-opencl                   Disable OpenCL GPU acceleration (default enabled)
  --consistency=2               Set depth map consistency filter level 0..2 (default 2)
  --denoise=1.0                 Set image denoise level (default 1.0)
  --save-steps                  Save intermediate images from processing steps
  --verbose                     Verbose output from steps
  --version                     Show application version number
  --opencv-version              Show OpenCV library version and build info

On Windows you can additionally just select the photos and drag them over focus-stack.exe to run with default settings.

On Mac OS X, you can start the program directly and it will ask for files. Alternatively you can drag the files over focus-stack application. If you want to provide command line parameters, open terminal and call the binary with path focus-stack.app/Contents/MacOS/focus-stack.

Algorithms used

The focus stacking algorithm used was invented and first described in Complex Wavelets for Extended Depth-of-Field: A New Method for the Fusion of Multichannel Microscopy Images by B. Forster, D. Van De Ville, J. Berent, D. Sage and M. Unser.

The application also uses multiple algorithms from OpenCV library. Most importantly, findTransformECC is used to align the source images.

Building on Ubuntu

The only dependency is OpenCV, 3.0 or newer, plus the basic build tools:

sudo apt install libopencv-dev build-essential

To enable GPU acceleration, you additionally need OpenCL library and GPU-specific driver:

sudo apt install ocl-icd-opencl-dev beignet            # For Intel GPUs
sudo apt install ocl-icd-opencl-dev nvidia-opencl-icd  # For NVidia GPUs

To build and install the application, simply type:

make
make install

Or to build a Debian/Ubuntu package and install it, type:

sudo apt install debhelper devscripts
make builddeb
sudo dpkg -i DEBUILD/focus-stack*.deb

Building on Windows

Download OpenCV binary package for Windows from OpenCV website. Download Build Tools for Visual Studio 2019 from Microsoft.

In Visual Studio command prompt, execute:

nmake -f Makefile.windows

Building on Mac OS X

Install Homebrew by running this in terminal:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install.sh)"

Then install build dependencies:

brew install opencv pkg-config dylibbundler

Finally, build and install to /usr/local/bin:

make
make install

Alternatively you can build an application bundle that can be used on any machine and that also includes a simple GUI for selecting the files:

make build/focus-stack.app