Skip to content

This repository demonstrates how to incorporate and use deep-learning based local features and matchers in COLMAP. Compared to the default SIFT features of COLMAP, it can provide denser and higher quality image matching.

License

Notifications You must be signed in to change notification settings

tmyok/custom-colmap

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Custom COLMAP

This repository demonstrates how to incorporate and use deep-learning based local features and matchers in COLMAP. Compared to the default SIFT features of COLMAP, it can provide denser and higher quality image matching.

Usage

convert.py is for preparing input image data for 3D Gaussian Splatting. Place the images you wish to use in the <location>/input directory.

<location>
|---input
    |---<image 0>
    |---<image 1>
    |---...

To run the Custom COLMAP code, which serves as an alternative to the original implementation of convert.py, execute the following command:

python3 convert.py --source_path <location>

Docker

Run the Docker container (if necessary).

sh docker_container.sh

About

This repository demonstrates how to incorporate and use deep-learning based local features and matchers in COLMAP. Compared to the default SIFT features of COLMAP, it can provide denser and higher quality image matching.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published