PhenoTypic — third-party component notices
==========================================

PhenoTypic is licensed under Apache-2.0 (see LICENSE). It optionally depends on
the third-party GPU-detection components below. PhenoTypic does NOT redistribute
any model weights: model weights are downloaded by the user from the upstream
source under that model's own license, which the user must accept.

- SAM2 (segment-anything-2), Meta — code & weights under Apache-2.0.
  Upstream:  https://github.com/facebookresearch/sam2
  License:   licenses/sam2-Apache-2.0.txt

- micro-sam (computational-cell-analytics) — code under the MIT License;
  checkpoints are distributed via its own channels under their respective terms.
  Upstream:  https://github.com/computational-cell-analytics/micro-sam
  License:   licenses/micro-sam-LICENSE.txt

- SAM3 (segment-anything-3), Meta — code via `transformers` (Apache-2.0);
  weights under the SAM License (commercial-OK, GATED on Hugging Face).
  Upstream:  https://huggingface.co/facebook/sam3
  License:   licenses/sam3-SAM-License.txt

- DINOv2, Meta — backbone for DinoSam2Detector; Apache-2.0, ungated.
  Upstream:  https://github.com/facebookresearch/dinov2
  License:   licenses/dinov2-Apache-2.0.txt

- DINOv3, Meta — frozen backbone for Insid3Detector (native) and the optional
  DINOv3 path of DinoSam2Detector / FssDinoDetector; weights under the DINOv3
  License (commercial-OK, redistributable only under that license, GATED on
  Hugging Face). Per the DINOv3 License §1.b.i, PhenoTypic prominently displays
  "Built with DINOv3" wherever DINOv3 features are produced.
  Upstream:  https://huggingface.co/facebook/dinov3-vitb16-pretrain-lvd1689m
  License:   licenses/dinov3-License.txt

- INSID3 (visinf/INSID3) — training-free in-context (one-shot) semantic
  segmentation on a frozen DINOv3 backbone; code under Apache-2.0. PhenoTypic's
  Insid3Detector is a clean-room reimplementation of its positional-bias-removal
  + prototype-matching method, attributed here; no upstream code is vendored.
  Upstream:  https://github.com/visinf/INSID3
  License:   licenses/insid3-Apache-2.0.txt

- FSSDINO (Zakir & Ho, "Revealing the Semantic Selection Gap in DINOv3 through
  Training-Free Few-Shot Segmentation", arXiv:2602.07550) — few-shot semantic
  segmentation (class-specific prototypes + Gram-matrix refinement) on a frozen
  DINO backbone. The reference code repository carries NO license (all rights
  reserved) and is NOT vendored; PhenoTypic's FssDinoDetector is a CLEAN-ROOM
  reimplementation FROM THE PAPER ONLY. The paper is licensed CC BY-NC-SA and is
  attributed here. No FSSDINO code is bundled or redistributed.
  Paper:     https://arxiv.org/abs/2602.07550

Reviewed filamentary-structure method ports
--------------------------------------------

The following notices cover the source-derived numerical cores and optional
runtime adapters introduced by the fungi detection method ports. Reference
papers, executable-source snapshots, oracle harnesses, and generated fixtures
under `docs/superpowers/` and `tests/fixtures/` are development evidence and
are excluded from wheel and source-distribution artifacts. Only production
code and the license texts listed below are distributed.

- Vaa3D APP2 grey-weighted distance transform, Vaa3D contributors. The
  `grey_weighted_distance` and `app2_gwdt_cost` implementations are derived
  from the pinned Vaa3D APP2 logic under the MIT License.
  Upstream:  https://github.com/Vaa3D/vaa3d_tools
  License:   licenses/vaa3d-MIT.txt

- Tensor Voting 2D Framework, Trevor Linton, copyright 2008. The upstream
  archive contains conflicting license statements: `license.txt` contains a
  BSD-like grant while `README` states LGPL-3.0 and includes that license.
  PhenoTypic preserves both claims without selecting between them. The
  source-derived `tensor_vote` component requires human licensing approval
  before a public release.
  Licenses:  licenses/tensor-voting-BSD-like.txt
             licenses/tensor-voting-LGPL-3.0-claim.txt

- TrickTrack, HSF contributors. The source-derived caller-supplied cellular
  automaton core is covered by Apache-2.0. No TrickTrack runtime dependency or
  upstream source is bundled.
  Upstream:  https://github.com/HSF/TrickTrack
  License:   licenses/tricktrack-Apache-2.0.txt

- LSD NFA scoring. The executable oracle is AGPL-3.0-or-later and is not
  distributed or linked. The production `binomial_nfa` core was independently
  authored in a source-free clean room from the published equation,
  source-generated fixtures, and a behavioral contract. Human licensing review
  of that clean-room boundary remains required before a public release.
  Paper:     https://doi.org/10.5201/ipol.2012.gjmr-lsd

- RORPO. The executable oracle is GPL-3.0-or-later and is not distributed or
  linked. The production `rorpo` core was independently authored in a
  source-free clean room from the IPOL paper and source-free generated
  fixtures. The paper and executable software have separate license terms.
  Paper:     https://doi.org/10.5201/ipol.2017.207

- Rolling Hough Transform, S. E. Clark. The Clark numerical core is derived
  from the pinned `seclark/RHT` implementation under the MIT License.
  Upstream:  https://github.com/seclark/RHT
  License:   licenses/rolling-hough-MIT.txt

- FilFinder, Eric Koch and Erik Rosolowsky, copyright 2014-2020. The optional
  `FilFinderDetector` adapter calls the separately installed `fil-finder==1.8`
  runtime under the MIT License. No FilFinder source is bundled.
  Upstream:  https://github.com/e-koch/FilFinder
  License:   licenses/filfinder-MIT.txt

- GUDHI, the GUDHI developers, copyright 2014. The optional
  `cubical_persistence` analysis calls the separately installed `gudhi==3.13.0`
  runtime under the MIT License. No GUDHI source or wheel is bundled.
  Upstream:  https://github.com/GUDHI/gudhi-devel
  License:   licenses/gudhi-MIT.txt

Gated foundation-model weights (SAM3, DINOv3) are never bundled; their downloads
are gated behind an explicit license-acceptance step (see
PHENOTYPIC_ACCEPT_MODEL_LICENSE and the require_license_acceptance hook in
src/phenotypic/detect/nn/_checkpoint_manager.py).
