#vision #descriptor #computer-vision #slam #sfm #keypoint

akaze

AKAZE feature extraction algorithm for computer vision

11 releases (6 breaking)

0.7.0 Jun 23, 2020
0.5.0 Jun 3, 2020
0.1.3 May 13, 2019
0.1.1 Mar 4, 2019
0.1.0 Oct 13, 2018

#4 in #sfm


Used in 2 crates

MIT license

1MB
1K SLoC

akaze

Discord Crates.io MIT/Apache docs.rs LoC Tests Lints

AKAZE feature extraction algorithm for computer vision

Implementation of AKAZE based on the one by indianajohn. He gave me permission to copy this here and work from that, as his job conflicts with maintainership. The crate is greatly changed from the original.

See tests/estimate_pose.rs for a demonstration on how to use this crate.

This crate adds several optimizations (using ndarray) to the original implementation and integrates directly into the rust-cv ecosystem for ease-of-use. This crate does not currently use threading to speed anything up, but it might be added as a Cargo feature in the future.

The original implementation can be found here: https://github.com/pablofdezalc/akaze

The previous rust implementation can be found here: https://github.com/indianajohn/akaze-rust

The main site for the algorithm is normally here, but it was down, so I found another link to the paper: http://www.bmva.org/bmvc/2013/Papers/paper0013/paper0013.pdf

Dependencies

~8.5MB
~162K SLoC