NOTICE

MegaDetector-Overhead
Copyright (c) 2024 Microsoft Corporation

Licensed under the MIT License — see LICENSE.txt at the repo root.

================================================================================
THIRD-PARTY SOFTWARE NOTICES
================================================================================

This repository vendors source code from two external projects. Each retains
its original license. The terms below apply only to the files inside the
named subdirectory; the rest of the repository is governed by the top-level
MIT LICENSE.txt.

--------------------------------------------------------------------------------
1. animaloc/  -- vendored from HerdNet
--------------------------------------------------------------------------------

Upstream:    https://github.com/Alexandre-Delplanque/HerdNet
Authors:     Alexandre Delplanque, Samuel Foucher, Jérôme Théau,
             Elsa Bussière, Cédric Vermeulen, Philippe Lejeune
             (Université de Liège, Gembloux Agro-Bio Tech, Forest Is Life)
License:     MIT License (compatible with the top-level LICENSE.txt)

Source paper:
    Delplanque, A., Foucher, S., Théau, J., Bussière, E., Vermeulen, C.,
    & Lejeune, P. (2023). From crowd to herd counting: How to precisely
    detect and count African mammals using aerial imagery and deep
    learning? ISPRS Journal of Photogrammetry and Remote Sensing, 197,
    167-180. https://doi.org/10.1016/j.isprsjprs.2023.01.025

The following files were renamed for the MegaDetector-Overhead release and
are described in docs/training.md:
    animaloc/models/owl_c.py  (was herdnet_detection_branch.py)
    animaloc/models/owl_d.py  (was herdnet_dinov2.py)
    animaloc/models/owl_t.py  (was herdnet_hybrid_multiscale_residual.py)

All other animaloc/ files are vendored as-is and retain their original MIT
copyright headers.

--------------------------------------------------------------------------------
2. dinov3/  -- vendored from facebookresearch/dinov3
--------------------------------------------------------------------------------

Upstream:    https://github.com/facebookresearch/dinov3
Authors:     Meta AI
License:     DINOv3 License (Meta Platforms, Inc.) -- see dinov3/LICENSE.md

NOTE: The DINOv3 License is NOT the MIT License. It is a custom Meta
license that permits use, reproduction, and modification for research and
commercial purposes subject to the terms in dinov3/LICENSE.md, including
attribution requirements and a redistribution clause.

Anyone redistributing the dinov3/ subdirectory (or works derived from it)
MUST include a copy of dinov3/LICENSE.md per section 1.b.i of that license.

Pretrained DINOv3 weights are NOT included in this repository (5.8 GB,
subject to the same DINOv3 License). Download instructions are in
INSTALL.md.

--------------------------------------------------------------------------------
Additional Python dependencies
--------------------------------------------------------------------------------

All transitive Python dependencies are listed in pyproject.toml with their
pinned versions in uv.lock. Each retains its own upstream license; see the
respective project pages for details.