{
  "name": "neuro-sam",
  "display_name": "Neuro-SAM",
  "visibility": "public",
  "icon": "",
  "categories": [],
  "schema_version": "0.2.1",
  "on_activate": null,
  "on_deactivate": null,
  "contributions": {
    "commands": [
      {
        "id": "neuro-sam.make_widget",
        "title": "Neuro-SAM",
        "python_name": "neuro_sam.napari_utils.main_widget:NeuroSAMWidget",
        "short_title": null,
        "category": null,
        "icon": null,
        "enablement": null
      }
    ],
    "readers": null,
    "writers": null,
    "widgets": [
      {
        "command": "neuro-sam.make_widget",
        "display_name": "Neuro-SAM",
        "autogenerate": false
      }
    ],
    "sample_data": null,
    "themes": null,
    "menus": {},
    "submenus": null,
    "keybindings": null,
    "configuration": []
  },
  "package_metadata": {
    "metadata_version": "2.4",
    "name": "neuro-sam",
    "version": "0.1.18",
    "dynamic": null,
    "platform": null,
    "supported_platform": null,
    "summary": "Neuro-SAM: Foundation Model for Dendrite and Dendritic Spine Segmentation",
    "description": "<div align=\"center\">\n\n# Neuro-SAM\n#### Foundation Model for Dendrite and Dendritic Spine Segmentation\n\n</div>\n\nNeuro-SAM is a napari plugin for analysing dendritic morphology in 3D fluorescence microscopy stacks. It enables you to:\n\n- Trace individual dendrites in a 3D volume using waypoint-guided brightest-path search\n- Segment traced dendrites using a fine-tuned SAM2 model\n- Detect dendritic spines along each traced path\n- Segment individual spines using a dedicated fine-tuned SAM2 model\n\nNeuro-SAM works across imaging modalities including two-photon, confocal, and STED microscopy.\n\n---\n\n### 🚀 Installation\n\nNeuro-SAM requires **Python 3.10+**. We recommend using Conda/Miniconda for environment management.\n\n```bash\npip install neuro-sam\n```\n\nModels are downloaded automatically on first use and stored in `src/neuro_sam/checkpoints/` — easy to find and delete.\n\nTo pre-download all models and the sample dataset:\n\n```bash\nneuro-sam-download\n```\n\nGPU acceleration is supported via **CUDA** and **MPS** (Apple Silicon).\n\n---\n\n### 📊 Usage\n\n```bash\n# Launch with the built-in benchmark dataset\nneuro-sam\n\n# Launch with your own dataset\nneuro-sam --image-path /path/to/your/image.tif\n```\n\n---\n\n### 🗃 Repository Structure\n\n```\nsrc/neuro_sam/\n├── checkpoints/          # model weights (auto-downloaded on first use)\n├── napari_utils/         # UI widgets and inference modules\n│   ├── main_widget.py          # top-level plugin widget\n│   ├── path_tracing_module.py  # waypoint path tracing\n│   ├── segmentation_module.py  # dendrite segmentation\n│   ├── segmentation_model.py   # SAM2 dendrite model wrapper\n│   ├── spine_detection_module.py     # spine detection\n│   ├── spine_segmentation_module.py  # spine segmentation UI\n│   ├── spine_segmentation_model.py   # SAM2 spine model wrapper\n│   └── anisotropic_scaling.py  # voxel spacing / scaling\n├── brightest_path_lib/   # waypoint A* path tracing algorithm\n├── training/             # training scripts for fine-tuning\n├── plugin.py             # napari entry point\n└── utils.py              # model download and shared utilities\n```\n\n---\n\n### 🔬 Workflow\n\n#### 1. Configure Voxel Spacing\nIn the **Path Tracing** tab, set accurate X, Y, Z voxel spacing (nm) for your acquisition. This ensures correct anisotropic scaling throughout the pipeline.\n\n#### 2. Trace Dendritic Paths\n- Click waypoints along dendrite structures in the viewer\n- Use **Load Dataset** to switch datasets without restarting\n- The algorithm finds the optimal brightest path between waypoints\n\n#### 3. Segment Dendrites\n- Go to the **Segmentation** tab\n- Load the pre-trained SAM2 dendrite model\n- Run segmentation on individual traced paths\n\n#### 4. Detect Spines\n- Go to the **Spine Detection** tab\n- Select a segmented path\n- Run automatic spine detection using tubular view analysis\n\n#### 5. Segment Spines\n- Go to the **Spine Segmentation** tab\n- Load the pre-trained SAM2 spine model\n- Run spine segmentation on detected spine positions\n- Export spine masks as TIFF files\n\n---\n\n### 🗂 Models\n\n| Model | Description |\n|---|---|\n| `sam2.1_hiera_small.pt` | SAM2 base checkpoint |\n| `dendrite_model.torch` | Fine-tuned weights for dendrite segmentation |\n| `spine_model.torch` | Fine-tuned weights for spine segmentation |\n\nAll models are downloaded from the [GitHub release](https://github.com/nipunarora8/Neuro-SAM/releases/tag/weights) on first use.\n\n---\n\n### 📬 Contact\n\n- Nipun Arora — nipunarora8@yahoo.com\n\n<div align=\"center\">\n<b>Made with ♥️ at <a href='https://anki.xyz'>Anki Lab</a> 🧠✨</b>\n</div>\n",
    "description_content_type": "text/markdown",
    "keywords": null,
    "home_page": null,
    "download_url": null,
    "author": null,
    "author_email": "Nipun Arora <nipunarora8@yahoo.com>",
    "maintainer": null,
    "maintainer_email": null,
    "license": "MIT",
    "classifier": [
      "Programming Language :: Python :: 3",
      "License :: OSI Approved :: MIT License",
      "Operating System :: OS Independent",
      "Framework :: napari"
    ],
    "requires_dist": [
      "napari",
      "numpy",
      "scipy",
      "imageio",
      "torch>=2.0.0",
      "torchvision>=0.15.0",
      "hydra-core>=1.3.2",
      "iopath>=0.1.10",
      "pillow>=9.4.0",
      "tqdm>=4.66.1",
      "vispy",
      "qtpy",
      "superqt",
      "magicgui",
      "scikit-image",
      "tifffile",
      "numba",
      "PyQt5",
      "opencv-python-headless",
      "matplotlib",
      "requests",
      "flammkuchen",
      "albumentations",
      "wandb",
      "tensorflow",
      "imagecodecs"
    ],
    "requires_python": ">=3.10",
    "requires_external": null,
    "project_url": [
      "Homepage, https://github.com/nipunarora8/Neuro-SAM",
      "Bug Tracker, https://github.com/nipunarora8/Neuro-SAM/issues"
    ],
    "provides_extra": null,
    "provides_dist": null,
    "obsoletes_dist": null
  },
  "npe1_shim": false
}