VisualGeometryToolkitFeatures detects, describes and matches image features: the ASIFT detector on the CPU or a GPU, nearest-neighbour descriptor matching, and the learned LoMa descriptors and matchers. It returns the geometric types of VisualGeometryToolkitCore.
pkg> add VisualGeometryToolkitFeaturesJulia 1.13 or later. Every dependency is in the General registry.
The example detects the blobs of a synthetic image and matches two sets of descriptors. It needs no files and no network.
using VisualGeometryToolkitFeatures
# Random is a standard library, shipped with Julia.
using Random
# A 120×160 gray image with three Gaussian blobs, given as
# (row, column, σ, amplitude): two light blobs and one dark one.
blobs = [(40, 50, 4.0, 0.35), (70, 100, 3.0, -0.3), (30, 130, 6.0, 0.3)]
img = [0.5 + sum(a * exp(-((r - y)^2 + (c - x)^2) / (2s^2)) for (y, x, s, a) in blobs)
for r in 1:120, c in 1:160]
# Detect with the default IMAS-25 tilt schedule.
dets = asift_detect(img)
length(dets) == 3 # true
count(is_dark, dets) == 1 # true
# A centre is (x, y) in COLMAP pixel coordinates, where the top-left pixel's
# centre is (0.5, 0.5): the blob at row 70, column 100 is found at (99.5, 69.5).
dark = only(filter(is_dark, dets))
round.(dark.center; digits = 1) == [99.5, 69.5] # true
# Match descriptors, one per row. Here they are random unit vectors, and the
# second set is the first shuffled and perturbed.
rng = Xoshiro(1)
unit_rows(D) = D ./ sqrt.(sum(abs2, D; dims = 2))
D1 = unit_rows(randn(rng, Float32, 100, 32))
perm = randperm(rng, 100)
D2 = unit_rows(D1[perm, :] .+ 0.05f0 .* randn(rng, Float32, 100, 32))
matches = nn_match(D1, D2) # mutual nearest neighbours, cosine similarity
all(perm[m.second] == m.first for m in matches) # trueThe Getting Started guide walks through it step by step.
- Getting Started: the example, step by step.
- Detection: the ASIFT detector, its coordinates, accuracy and detection covariance, and a lens's image circle measured from photographs.
- Matching: strategies, metrics, stacked galleries and LoMa.
- API Reference.
VisualGeometryToolkitIO reads and writes detections as JSON, and VisualGeometryToolkitEstimation fits models to the matches. The portal links every package of the toolkit.
VisualGeometryToolkitFeatures is part of VisualGeometryToolkit. Cite the toolkit. Its CITATION.cff is the source of this entry, and GitHub's "Cite this repository" button exports it:
@software{pritts2026vgtk,
author = {Pritts, James and Sauer, Hendrik},
title = {{VisualGeometryToolkit.jl}: Julia Packages for Multiple-View Geometry},
url = {https://github.com/VisualGeometryToolkit/VisualGeometryToolkit.jl},
doi = {10.5281/zenodo.23548372},
year = {2026}
}Cite also the papers behind the methods you use, listed on the portal.
LoMaDescriptor and LoMaMatcher run the networks of
LoMa by David Nordström, Johan Edstedt,
Georg Bökman, Jonathan Astermark, Anders Heyden, Viktor Larsson, Mårten
Wadenbäck, Michael Felsberg and Fredrik Kahl. The architectures, the trained
weights and the ONNX exports are theirs, downloaded from the authors' release.
The descriptors are DeDoDe's (Edstedt
et al.). If you use them in published work, cite both:
@inproceedings{nordstrom2026loma,
title = {{LoMa: Local Feature Matching Revisited}},
author = {David Nordström and Johan Edstedt and Georg Bökman and Jonathan Astermark and Anders Heyden and Viktor Larsson and Mårten Wadenbäck and Michael Felsberg and Fredrik Kahl},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2026}
}
@inproceedings{edstedt2024dedode,
title = {{DeDoDe: Detect, Don't Describe --- Describe, Don't Detect for Local Feature Matching}},
author = {Johan Edstedt and Georg Bökman and Mårten Wadenbäck and Michael Felsberg},
booktitle = {2024 International Conference on 3D Vision (3DV)},
year = {2024},
organization = {IEEE}
}LoMa's code is MIT-licensed except its matcher, which inherits LightGlue's Apache-2.0; DeDoDe is MIT except the DINOv2 backbone of DeDoDe-G (Apache-2.0).
spread_strongest is Suppression via Square Covering (SSC) by Oleksandr Bailo,
Francois Rameau, Kyungdon Joo, Jinsun Park, Oleksandr Bogdan and In So Kweon,
ported from the authors' code (ANMS-Codes,
MIT License, Copyright (c) 2018 Oleksandr Bailo) and parallelised on
KernelAbstractions. If you use it in published work, cite:
@article{bailo2018efficient,
title = {Efficient adaptive non-maximal suppression algorithms for homogeneous spatial keypoint distribution},
author = {Bailo, Oleksandr and Rameau, Francois and Joo, Kyungdon and Park, Jinsun and Bogdan, Oleksandr and Kweon, In So},
journal = {Pattern Recognition Letters},
volume = {106},
pages = {53--60},
year = {2018}
}MIT