DRerio LogAI: an open-source, GPU-free platform for automated tracking and behavioural analysis of adult zebrafish (Danio rerio)

DRerio LogAI is an open-source desktop platform for automated tracking, closed-loop stimulation and behavioural analysis of adult zebrafish (Danio rerio). It combines YOLOv11 detection, BYTETracker multi-object tracking and Intel OpenVINO inference in an event-driven architecture, and runs GPU-free on integrated graphics.

This archive. Snapshot of the source code at tag v7.0.0, deposited as the citable code object for the associated manuscripts. The validation and benchmark results reported in those manuscripts were produced with release 4.0.0, which remains available as tag v4.0.0 for exact reproduction; releases 5, 6 and 7 optimised the detector plugins, the user interface and internationalisation, and refined the closed-loop session logger. The analysis scripts and the paired coordinate tables that regenerate every reported statistic are deposited separately (see related identifiers).

Licensing. The original source code of this repository is licensed MIT, copyright Universidade Estadual Paulista (UNESP). The application as distributed is effectively licensed AGPL-3.0-or-later, because it packages and requires Ultralytics YOLO (AGPL-3.0). The trained YOLOv11 weights are not part of this source archive: they are published as assets of a separate GitHub release and installed by the bundled fetch-weights command, which verifies each file against a SHA-256 recorded in weights_manifest.json. Those weights are Ultralytics derivatives and are likewise effectively AGPL-3.0; their training imagery derives from the ZebraFish-Detection dataset (Roboflow Universe), licensed CC BY 4.0, which requires attribution. Companion hardware designs are licensed CERN-OHL-S v2. See the LICENSE and NOTICE files in the archive.

Intellectual property. Registered as a computer program with the Brazilian National Institute of Industrial Property (INPI), process BR 51 2026 005215-7 (petition 870260066857, filed 2026-07-07). Patrimonial rights are held by Universidade Estadual Paulista (UNESP), CNPJ 48.031.918/0001-24; moral rights by the authors. This is a copyright deposit under Brazilian Law 9.609/98 and does not restrict the open-source licensing above. Public release authorised by the UNESP technology transfer office (AUIN).

Funding. São Paulo Research Foundation (FAPESP), grant 2023/14200-3. Animal procedures approved by the UNESP Animal Use Ethics Committee (CEUA), protocol 4806060624.

Requirements. Python ≥ 3.12, < 3.15; Ultralytics YOLO ≥ 8.3.179; Intel OpenVINO ~2026.0; PyTorch ≥ 2.8; NumPy ~2.2; OpenCV ~4.13. Full dependency pins in pyproject.toml and poetry.lock. Validated on Windows 11; platform-independent Python otherwise.