{"id":31684743,"url":"https://github.com/mjahmadee/cataract-lmm","last_synced_at":"2026-05-15T08:41:06.640Z","repository":{"id":313893155,"uuid":"1053331476","full_name":"MJAHMADEE/Cataract-LMM","owner":"MJAHMADEE","description":"Advanced computer vision framework for surgical video analysis, featuring multi-task learning for cataract surgery procedure understanding, instrument segmentation, and skill assessment.","archived":false,"fork":false,"pushed_at":"2025-09-27T07:08:01.000Z","size":2162,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-09-27T09:18:30.422Z","etag":null,"topics":["ai-research","cataract-surgery","computer-vision","deep-learning","healthcare","instance-segmentation","machine-learning","medical-ai","medical-imaging","phase-recognition","pytorch","skill-assessment","surgical-analysis","surgical-robotics","video-analysis","video-processing"],"latest_commit_sha":null,"homepage":"https://github.com/MJAHMADEE/Cataract-LMM","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/MJAHMADEE.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":".github/CODEOWNERS","security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-09-09T09:53:22.000Z","updated_at":"2025-09-14T18:06:42.000Z","dependencies_parsed_at":"2025-09-09T12:39:29.990Z","dependency_job_id":"dfe96544-3453-44b9-b51b-7adc5b3c1a44","html_url":"https://github.com/MJAHMADEE/Cataract-LMM","commit_stats":null,"previous_names":["mjahmadee/cataract-lmm"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/MJAHMADEE/Cataract-LMM","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MJAHMADEE%2FCataract-LMM","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MJAHMADEE%2FCataract-LMM/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MJAHMADEE%2FCataract-LMM/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MJAHMADEE%2FCataract-LMM/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/MJAHMADEE","download_url":"https://codeload.github.com/MJAHMADEE/Cataract-LMM/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MJAHMADEE%2FCataract-LMM/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":278915788,"owners_count":26068085,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-10-08T02:00:06.501Z","response_time":56,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-research","cataract-surgery","computer-vision","deep-learning","healthcare","instance-segmentation","machine-learning","medical-ai","medical-imaging","phase-recognition","pytorch","skill-assessment","surgical-analysis","surgical-robotics","video-analysis","video-processing"],"created_at":"2025-10-08T08:38:32.300Z","updated_at":"2026-05-15T08:41:06.633Z","avatar_url":"https://github.com/MJAHMADEE.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🏥 Cataract-LMM: Surgical Video Analysis\n\n[![Python 3.8+](https://img.shields.io/badge/Python-3.8%2B-3776ab.svg?style=flat\u0026logo=python\u0026logoColor=white)](https://python.org)\n[![PyTorch](https://img.shields.io/badge/PyTorch-2.0%2B-ee4c2c.svg?style=flat\u0026logo=pytorch\u0026logoColor=white)](https://pytorch.org)\n[![Poetry](https://img.shields.io/badge/Dependency%20Management-Poetry-60a5fa.svg?style=flat\u0026logo=poetry\u0026logoColor=white)](https://python-poetry.org)\n[![Docker](https://img.shields.io/badge/Container-Docker-2496ed.svg?style=flat\u0026logo=docker\u0026logoColor=white)](https://docker.com)\n[![License: CC-BY-4.0](https://img.shields.io/badge/License-CC--BY--4.0-lightgrey.svg?style=flat)](https://creativecommons.org/licenses/by/4.0/)\n[![CI/CD](https://img.shields.io/badge/CI%2FCD-GitHub%20Actions-2088ff.svg?style=flat\u0026logo=github-actions\u0026logoColor=white)](https://github.com/features/actions)\n[![Code Quality](https://img.shields.io/badge/Code%20Quality-Black%20%7C%20Flake8%20%7C%20MyPy-000000.svg?style=flat\u0026logo=python\u0026logoColor=white)](https://github.com/psf/black)\n[![Medical AI](https://img.shields.io/badge/Domain-Medical%20AI-26a69a.svg?style=flat\u0026logo=stethoscope\u0026logoColor=white)](#)\n\n**A comprehensive, production-ready framework for multi-task deep learning in surgical video analysis, featuring instance segmentation, phase recognition, skill assessment, and video processing capabilities.**\n\n---\n\n## 🎯 Overview\n\nCataract-LMM is an enterprise-grade AI framework designed for large-scale, multi-center surgical video analysis. Built on modern software engineering principles, this repository provides state-of-the-art deep learning models for comprehensive analysis of cataract surgery videos.\n\n### 🔬 Research Foundation\nThis framework implements methodologies from cutting-edge research in computer-assisted surgery, providing validated approaches for:\n- **Surgical Instance Segmentation** using YOLO, Mask R-CNN, and SAM architectures\n- **Surgical Phase Recognition** with Video Transformers, 3D CNNs, and temporal models  \n- **Surgical Skill Assessment** through multi-modal analysis and performance metrics\n- **Video Processing** with GPU-accelerated pipelines for medical video data\n\n### 🏆 Key Differentiators\n- **Production-Ready**: Enterprise-grade architecture with comprehensive testing and CI/CD\n- **Multi-Task Learning**: Unified framework supporting four core surgical analysis tasks\n- **Scalable Design**: Microservices-ready architecture with containerization support\n- **Medical Compliance**: HIPAA-aware design patterns and secure data handling\n- **Research-to-Production**: Seamless transition from research notebooks to production deployment\n\n---\n\n## 📋 Table of Contents\n\n- [🚀 Quick Start](#-quick-start)\n- [✨ Features](#-features)\n- [🏗️ Architecture](#️-architecture)\n- [📦 Installation](#-installation)\n- [🎯 Usage Examples](#-usage-examples)\n- [🛠️ Development](#️-development)\n- [📊 Model Zoo](#-model-zoo)\n- [🔧 Configuration](#-configuration)\n- [🧪 Testing](#-testing)\n- [📖 Documentation](#-documentation)\n- [🤝 Contributing](#-contributing)\n- [📄 License](#-license)\n- [📣 Citation](#-citation)\n- [👨‍💻 Author](#-author)\n- [📞 Support \u0026 Community](#-support--community)\n- [🚀 Roadmap](#-roadmap)\n\n---\n\n## 🚀 Quick Start\n\n### Prerequisites\n- Python 3.8+\n- CUDA 11.8+ (for GPU acceleration)\n- FFmpeg (for video processing)\n- Docker (optional, for containerized deployment)\n\n### Installation\n\n```bash\n# Clone the repository\ngit clone https://github.com/MJAHMADEE/Cataract_LMM.git\ncd Cataract_LMM\n\n# Install using Poetry (recommended)\ncd codes\npoetry install\n\n# Activate virtual environment\npoetry shell\n\n# Or install using pip\npip install -r requirements.txt\n\n# Validate installation\npython setup.py --validate-only\n```\n\n### Basic Usage\n\n```bash\n# Video processing\ncd surgical-video-processing\npython main.py --input path/to/video.mp4 --output ./results --config configs/default.yaml\n\n# Instance segmentation  \ncd surgical-instance-segmentation\npython inference/predictor.py --model yolo --input data/images/\n\n# Phase recognition\ncd surgical-phase-recognition\npython validation/training_framework.py --config configs/default.yaml --mode train\n\n# Skill assessment\ncd surgical-skill-assessment\npython main.py --config configs/comprehensive.yaml --mode evaluate\n```\n\n---\n\n## ✨ Features\n\n### 🧠 AI/ML Capabilities\n\n| Component | Models | Key Features |\n|-----------|--------|-------------|\n| **Instance Segmentation** | YOLO v8/11, Mask R-CNN, SAM | Real-time surgical instrument detection and segmentation |\n| **Phase Recognition** | Video Transformers, 3D CNNs, TeCNO | 11-phase surgical workflow analysis |\n| **Skill Assessment** | Multi-modal CNNs, Attention Models | Objective surgical skill evaluation |\n| **Video Processing** | GPU-Accelerated Pipelines | Medical-grade video preprocessing and enhancement |\n\n### 🛠️ Engineering Excellence\n\n- **🏗️ Modular Architecture**: Microservices-ready design with clear separation of concerns\n- **🔒 Security First**: HIPAA-compliant patterns, secure credential management\n- **📊 Comprehensive Testing**: 85%+ test coverage with unit, integration, and E2E tests\n- **🚀 CI/CD Pipeline**: Automated testing, security scanning, and deployment workflows\n- **📈 Monitoring \u0026 Observability**: Structured logging, metrics collection, and health checks\n- **🐳 Containerization**: Multi-stage Docker builds with security hardening\n\n### 🔧 Developer Experience\n\n- **📚 Rich Documentation**: Comprehensive guides, API references, and examples\n- **🎯 Configuration Management**: YAML-based configuration with validation\n- **🧪 Development Tools**: Pre-commit hooks, linting, formatting, and type checking\n- **📦 Dependency Management**: Poetry-based modern Python packaging\n- **🔧 Development Environment**: VS Code integration with debugging support\n\n---\n\n## 🏗️ Architecture\n\n### System Overview\n```mermaid\ngraph TB\n    A[Video Input] --\u003e B[Video Processing Pipeline]\n    B --\u003e C[Frame Extraction \u0026 Preprocessing]\n    C --\u003e D[Multi-Task Analysis Engine]\n    \n    D --\u003e E[Instance Segmentation]\n    D --\u003e F[Phase Recognition] \n    D --\u003e G[Skill Assessment]\n    \n    E --\u003e H[Surgical Instruments]\n    F --\u003e I[Surgery Phases]\n    G --\u003e J[Skill Metrics]\n    \n    H --\u003e K[Clinical Decision Support]\n    I --\u003e K\n    J --\u003e K\n```\n\n### Project Structure\n```\nCataract_LMM/\n├── 🏠 README.md                    # Project overview and documentation\n├── 📄 LICENSE                     # CC-BY-4.0 license\n├── 🤝 CONTRIBUTING.md             # Contribution guidelines\n├── 🔒 .gitignore                  # Git ignore patterns\n├── 📊 codes/                      # Main codebase\n│   ├── 🎬 surgical-video-processing/          # Video preprocessing and enhancement\n│   │   ├── core/                  # Core processing algorithms\n│   │   ├── pipelines/             # Processing pipelines\n│   │   ├── metadata/              # Video metadata management\n│   │   ├── quality_control/       # Quality assurance tools\n│   │   └── configs/               # Configuration files\n│   ├── 🎯 surgical-instance-segmentation/     # Instance segmentation models\n│   │   ├── models/                # YOLO, Mask R-CNN, SAM implementations\n│   │   ├── training/              # Training pipelines\n│   │   ├── inference/             # Real-time inference engines\n│   │   ├── evaluation/            # Model evaluation tools\n│   │   └── data/                  # Dataset utilities\n│   ├── 🔄 surgical-phase-recognition/         # Phase classification models\n│   │   ├── models/                # Video Transformers, 3D CNNs, TeCNO\n│   │   ├── validation/            # Training and validation frameworks\n│   │   ├── preprocessing/         # Video preprocessing\n│   │   ├── analysis/              # Result analysis tools\n│   │   └── configs/               # Model configurations\n│   ├── 📊 surgical-skill-assessment/          # Skill evaluation framework\n│   │   ├── models/                # Skill assessment models\n│   │   ├── engine/                # Training and inference engines\n│   │   ├── utils/                 # Analysis utilities\n│   │   └── configs/               # Assessment configurations\n│   ├── 🧪 tests/                  # Comprehensive test suite\n│   ├── 📚 docs/                   # Documentation source\n│   ├── 🐳 docker/                 # Docker configurations\n│   ├── 📊 reports/                # Analysis reports\n│   ├── ⚙️ pyproject.toml         # Python project configuration\n│   ├── 🔒 Dockerfile             # Container definition\n│   ├── 🚀 Makefile               # Development automation\n│   └── 🔧 setup.py               # Project setup script\n├── 🤖 .github/                   # GitHub configurations\n│   └── workflows/                 # CI/CD pipelines\n└── 📓 security_scanning_demo.ipynb  # Security analysis notebook\n```\n\n---\n\n## 📦 Installation\n\n### System Requirements\n\n| Component | Minimum | Recommended |\n|-----------|---------|-------------|\n| **Python** | 3.8 | 3.11+ |\n| **RAM** | 16GB | 32GB+ |\n| **GPU Memory** | 8GB | 24GB+ |\n| **Storage** | 50GB | 500GB+ |\n| **CUDA** | 11.8 | 12.0+ |\n\n### Installation Methods\n\n#### Method 1: Poetry (Recommended)\n```bash\n# Install Poetry\ncurl -sSL https://install.python-poetry.org | python3 -\n\n# Clone and setup\ngit clone https://github.com/MJAHMADEE/Cataract_LMM.git\ncd Cataract_LMM/codes\n\n# Install dependencies\npoetry install --extras \"dev docs\"\n\n# Activate environment\npoetry shell\n```\n\n#### Method 2: Conda Environment\n```bash\n# Create environment\nconda create -n cataract-lmm python=3.11\nconda activate cataract-lmm\n\n# Clone and install\ngit clone https://github.com/MJAHMADEE/Cataract_LMM.git\ncd Cataract_LMM/codes\npip install -r requirements.txt\n```\n\n#### Method 3: Docker Deployment\n```bash\n# Build container\ndocker build -t cataract-lmm:latest .\n\n# Run interactive container\ndocker run -it --gpus all -v $(pwd)/data:/app/data cataract-lmm:latest\n```\n\n### Verification\n```bash\n# Run comprehensive validation\npython setup.py --validate-only\n\n# Run tests\npytest tests/ -v\n\n# Check GPU availability\npython -c \"import torch; print(f'CUDA Available: {torch.cuda.is_available()}')\"\n```\n\n---\n\n## 🎯 Usage Examples\n\n### Video Processing Pipeline\n```python\nfrom surgical_video_processing import VideoProcessor, QualityController\n\n# Initialize processor with configuration\nprocessor = VideoProcessor(\"configs/high_quality.yaml\")\n\n# Process surgical video\nresult = processor.process_video(\n    input_path=\"data/surgery_video.mp4\",\n    output_dir=\"outputs/processed/\",\n    apply_deidentification=True,\n    quality_threshold=0.8\n)\n\nprint(f\"Processed {result.frame_count} frames\")\nprint(f\"Quality score: {result.average_quality:.3f}\")\n```\n\n### Instance Segmentation\n```python  \nfrom surgical_instance_segmentation import SegmentationPredictor\n\n# Load pre-trained model\npredictor = SegmentationPredictor(\n    model_type=\"yolo_v8\",\n    device=\"cuda\"\n)\n\n# Segment surgical instruments\nresults = predictor.predict_batch(\n    image_paths=[\"frame001.jpg\", \"frame002.jpg\"],\n    confidence_threshold=0.7,\n    save_visualizations=True\n)\n\n# Extract detections\nfor result in results:\n    print(f\"Detected {len(result.boxes)} instruments\")\n    print(f\"Classes: {result.class_names}\")\n```\n\n### Phase Recognition\n```python\nfrom surgical_phase_recognition import PhaseClassifier\n\n# Initialize phase recognition model\nclassifier = PhaseClassifier(\n    model_name=\"video_transformer\",\n    config_path=\"configs/phase_recognition.yaml\"\n)\n\n# Classify surgical phases in video sequence\nphases = classifier.classify_sequence(\n    video_path=\"data/surgery_complete.mp4\",\n    sequence_length=16,\n    overlap=0.5\n)\n\n# Display phase timeline\nfor phase in phases:\n    print(f\"Time: {phase.timestamp:.2f}s - Phase: {phase.name}\")\n```\n\n### Skill Assessment\n```python\nfrom surgical_skill_assessment import SkillEvaluator\n\n# Initialize skill assessment framework\nevaluator = SkillEvaluator(\"configs/skill_assessment.yaml\")\n\n# Assess surgical performance\nassessment = evaluator.evaluate_surgery(\n    video_path=\"data/complete_surgery.mp4\",\n    phase_annotations=\"data/phases.json\",\n    surgeon_level=\"resident\"  # resident, fellow, attending\n)\n\n# Generate skill report\nreport = evaluator.generate_report(assessment)\nprint(f\"Overall Score: {report.overall_score}/100\")\nprint(f\"Efficiency: {report.efficiency_score}/10\")\nprint(f\"Precision: {report.precision_score}/10\")\n```\n\n---\n\n## 🛠️ Development\n\n### Development Setup\n```bash\n# Clone repository\ngit clone https://github.com/MJAHMADEE/Cataract_LMM.git\ncd Cataract_LMM/codes\n\n# Install development dependencies\npoetry install --extras \"dev\"\n\n# Setup pre-commit hooks\npre-commit install\n\n# Run development server\nmake dev-server\n```\n\n### Code Quality Tools\n```bash\n# Format code\nmake format\n\n# Run linting\nmake lint\n\n# Type checking\nmake type-check\n\n# Security scanning\nmake security-scan\n\n# Run all quality checks\nmake quality\n```\n\n### Testing\n```bash\n# Run unit tests\nmake test\n\n# Run with coverage\nmake test-coverage\n\n# Run integration tests\nmake test-integration\n\n# Run end-to-end tests\nmake test-e2e\n\n# Generate coverage report\nmake coverage-report\n```\n\n### Available Make Commands\n```bash\nmake help              # Show all available commands\nmake install           # Install dependencies\nmake clean             # Clean build artifacts\nmake build             # Build distribution packages\nmake docker-build      # Build Docker image\nmake docker-run        # Run Docker container\nmake docs-build        # Build documentation\nmake docs-serve        # Serve documentation locally\n```\n\n---\n\n## 📊 Model Zoo\n\n### Instance Segmentation Models (Task 3: 12-class)\n\n| Model | mAP@0.5:0.95 | \n|-------|-------------|\n| **YOLOv11** ⭐ | **73.9%** |\n| **YOLOv8** | **73.8%** |\n| **SAM** | **56.0%** |\n| **SAM2** | **55.2%** |\n| **Mask R-CNN** | **53.7%** |\n\n### Phase Recognition Models (In-Domain - Farabi Test Set)\n\n| Model | Backbone | Accuracy | F1-Score | Precision | Recall |\n|-------|----------|----------|----------|-----------|---------|\n| **MViT-B** ⭐ | - | **85.7%** | **77.1%** | **77.1%** | **78.5%** |\n| **Swin-T** | - | **85.5%** | **76.2%** | **77.5%** | **77.2%** |\n| **CNN + GRU** | EfficientNet-B5 | **82.1%** | **71.3%** | **76.0%** | **70.4%** |\n| **CNN + TeCNO** | EfficientNet-B5 | **81.7%** | **71.2%** | **75.1%** | **71.2%** |\n| **CNN + LSTM** | EfficientNet-B5 | **81.5%** | **70.0%** | **76.4%** | **69.4%** |\n\n### Skill Assessment Models\n\n| Model | Accuracy | Precision | Recall | F1-Score |\n|-------|----------|-----------|---------|----------|\n| **TimeSformer** ⭐ | **82.5%** | **86.0%** | **82.0%** | **83.9%** |\n| **R3D-18** | **81.7%** | **82.4%** | **84.9%** | **83.6%** |\n| **Slow R50** | **80.0%** | **81.8%** | **81.8%** | **81.8%** |\n| **X3D-M** | **80.0%** | **83.9%** | **78.8%** | **81.3%** |\n| **R(2+1)D-18** | **72.9%** | **79.3%** | **76.7%** | **78.0%** |\n\n---\n\n## 🔧 Configuration\n\n### Configuration Files\n\nThe framework uses YAML-based configuration for all components:\n\n#### Video Processing (`surgical-video-processing/configs/default.yaml`)\n```yaml\nprocessing:\n  target_resolution: [1920, 1080]\n  fps: 30\n  quality_threshold: 0.75\n  \ndeidentification:\n  enabled: true\n  blur_faces: true\n  remove_text: true\n  \noutput:\n  format: \"mp4\"\n  compression: \"h264\"\n  quality: \"high\"\n```\n\n#### Instance Segmentation (`surgical-instance-segmentation/configs/yolo_config.yaml`)\n```yaml\nmodel:\n  architecture: \"yolov8\"\n  size: \"medium\"\n  pretrained: true\n\ntraining:\n  epochs: 100\n  batch_size: 16\n  learning_rate: 0.001\n  \ndata:\n  classes: [\"forceps\", \"scissors\", \"needle_holder\", \"suction\"]\n  augmentation:\n    enabled: true\n    rotation: 15\n    scaling: [0.8, 1.2]\n```\n\n### Environment Variables\n```bash\n# Create .env file\ncp .env.example .env\n\n# Edit configuration\nCUDA_VISIBLE_DEVICES=0,1\nWANDB_PROJECT=cataract-lmm\nDATA_ROOT=/path/to/data\nOUTPUT_DIR=/path/to/outputs\nLOG_LEVEL=INFO\n```\n\n---\n\n## 🧪 Testing\n\n### Test Structure\n```\ntests/\n├── unit/                   # Unit tests for individual components\n├── integration/            # Integration tests for module interactions  \n├── e2e/                   # End-to-end workflow tests\n├── performance/           # Performance and benchmarking tests\n├── security/              # Security and vulnerability tests\n├── fixtures/              # Test data and fixtures\n└── conftest.py           # Pytest configuration\n```\n\n### Running Tests\n```bash\n# Run all tests\npytest\n\n# Run specific test category\npytest tests/unit/\npytest tests/integration/\npytest tests/e2e/\n\n# Run with coverage\npytest --cov=. --cov-report=html\n\n# Run performance tests\npytest tests/performance/ --benchmark-only\n\n# Run with specific markers\npytest -m \"gpu\" --gpu-required\npytest -m \"slow\" --timeout=300\n```\n\n### Test Configuration\n```ini\n# pytest.ini\n[tool:pytest]\ntestpaths = tests\npython_files = test_*.py\npython_classes = Test*\npython_functions = test_*\nmarkers =\n    unit: Unit tests\n    integration: Integration tests\n    e2e: End-to-end tests\n    gpu: Tests requiring GPU\n    slow: Slow running tests\n    security: Security tests\naddopts = \n    --strict-markers\n    --verbose\n    --tb=short\n    --cov-report=term-missing\n```\n\n---\n\n## 📖 Documentation\n\n### Documentation Structure\n- **📚 User Guide**: Getting started, tutorials, and examples\n- **🔧 API Reference**: Comprehensive API documentation\n- **🏗️ Developer Guide**: Contributing, architecture, and development setup\n- **📊 Model Documentation**: Model architectures, performance metrics, and usage\n- **🔐 Security Guide**: Security considerations and best practices\n\n### Building Documentation\n```bash\n# Install documentation dependencies\npoetry install --extras \"docs\"\n\n# Build documentation\ncd docs\nmake html\n\n# Serve documentation locally\nmake serve\n\n# Build PDF documentation\nmake latexpdf\n```\n\n### Online Documentation\n- **Documentation Site**: https://cataract-lmm.readthedocs.io\n- **API Reference**: https://cataract-lmm.readthedocs.io/api/\n- **Tutorials**: https://cataract-lmm.readthedocs.io/tutorials/\n- **Model Zoo**: https://cataract-lmm.readthedocs.io/models/\n\n---\n\n## 🤝 Contributing\n\nWe welcome contributions from the surgical AI community! Please see our [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines.\n\n### Quick Contribution Guide\n1. **Fork** the repository\n2. **Create** a feature branch (`git checkout -b feature/amazing-feature`)\n3. **Commit** changes (`git commit -m 'Add amazing feature'`)\n4. **Push** to branch (`git push origin feature/amazing-feature`)\n5. **Open** a Pull Request\n\n### Development Workflow\n```bash\n# Setup development environment\nmake dev-setup\n\n# Run pre-commit checks\npre-commit run --all-files\n\n# Run tests before committing\nmake test-all\n\n# Submit pull request\ngh pr create --title \"Feature: Add amazing feature\"\n```\n\n### Code Standards\n- **Python Style**: [Black](https://black.readthedocs.io/) formatter\n- **Import Sorting**: [isort](https://pycqa.github.io/isort/)\n- **Linting**: [Flake8](https://flake8.pycqa.org/) with medical AI conventions\n- **Type Checking**: [MyPy](https://mypy.readthedocs.io/) for type safety\n- **Documentation**: [Google style](https://google.github.io/styleguide/pyguide.html) docstrings\n\n---\n\n## 📄 License\n\n### **Framework License**\nThis project framework and code are licensed under the **Creative Commons Attribution 4.0 International License** (CC-BY-4.0). See the [LICENSE](LICENSE) file for details.\n\n### **Data License**\nThe dataset has specific ownership and licensing requirements. See [DATA_LICENSE.md](DATA_LICENSE.md) for detailed information about:\n- Data ownership by Farabi Eye Hospital and Noor Eye Hospital\n- Annotation ownership by participating institutions  \n- Attribution requirements under CC-BY 4.0\n- Proper usage guidelines\n\n---\n\n## 📣 Citation\n\nIf you use this benchmark dataset or framework in your research, please cite our work. The benchmark has been submitted to *Scientific Data* (Nature Portfolio).\n\n### BibTeX\n```bibtex\n@misc{ahmadi2025cataractlmmlargescalemultitask,\n      title={Cataract-LMM: Large-Scale, Multi-Source, Multi-Task Benchmark for Deep Learning in Surgical Video Analysis}, \n      author={Mohammad Javad Ahmadi and Iman Gandomi and Parisa Abdi and Seyed-Farzad Mohammadi and Amirhossein Taslimi and Mehdi Khodaparast and Hassan Hashemi and Mahdi Tavakoli and Hamid D. Taghirad},\n      year={2025},\n      eprint={2510.16371},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2510.16371},\n      doi={10.48550/arXiv.2510.16371}\n}\n```\n\n### APA Style\nAhmadi, M. J., Gandomi, I., Abdi, P., Mohammadi, S.-F., Taslimi, A., Khodaparast, M., Hashemi, H., Tavakoli, M., \u0026 Taghirad, H. D. (2025). Cataract-LMM: Large-Scale, Multi-Source, Multi-Task Benchmark for Deep Learning in Surgical Video Analysis. arXiv. https://doi.org/10.48550/arXiv.2510.16371\n\n### IEEE Style\nM. J. Ahmadi et al., \"Cataract-LMM: Large-Scale, Multi-Source, Multi-Task Benchmark for Deep Learning in Surgical Video Analysis,\" 2025, arXiv:2510.16371. [Online]. Available: https://arxiv.org/abs/2510.16371\n\n### Chicago Style\nAhmadi, Mohammad Javad, Iman Gandomi, Parisa Abdi, Seyed-Farzad Mohammadi, Amirhossein Taslimi, Mehdi Khodaparast, Hassan Hashemi, Mahdi Tavakoli, and Hamid D. Taghirad. 2025. \"Cataract-LMM: Large-Scale, Multi-Source, Multi-Task Benchmark for Deep Learning in Surgical Video Analysis.\" arXiv. https://doi.org/10.48550/arXiv.2510.16371.\n\n### Repository Citation\n```bibtex\n@software{cataract_lmm_repo_2025,\n  title={{Cataract-LMM}: Large-Scale, Multi-Source, Multi-Task Benchmark and Framework for Surgical Video Analysis},\n  author={Ahmadi, Mohammad Javad and Gandomi, Iman and Abdi, Parisa and Mohammadi, Seyed-Farzad and Taslimi, Amirhossein and Khodaparast, Mehdi and Hashemi, Hassan and Tavakoli, Mahdi and Taghirad, Hamid D.},\n  year={2025},\n  url={https://github.com/MJAHMADEE/Cataract-LMM},\n  version={1.0.0}\n}\n```\n\n---\n\n## 👨‍💻 Author\n\n**Mohammad Javad Ahmadi**\n\n\u003cp\u003e\n\u003ca href=\"https://drive.google.com/file/d/12mfH_QxeocmGwL8oAGfFRh2LZSZSlrwE/view?usp=sharing\" target=\"_blank\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Resume-View%20CV-blue?style=for-the-badge\u0026logo=google-drive\" alt=\"Resume\" /\u003e\n\u003c/a\u003e\n\u003c/p\u003e\n\n\u003cp\u003e\n\u003ca href=\"http://mjahmadee.site123.me/\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454571884/website.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"mailto:mjahmadee@gmail.com\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454571876/email.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://www.linkedin.com/in/mjahmadi/\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454571918/LinkedIn.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://www.researchgate.net/profile/Mohammad_Ahmadi67\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454571968/RG.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://scholar.google.com/citations?user=wTnN9IEAAAAJ\u0026hl=en\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454571984/GScholar.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://orcid.org/0000-0002-2177-9936\" rel=\"noopener noreferrer\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454572050/orcid.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"http://www.webofscience.com/wos/author/record/GXN-0273-2022\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454572084/publons.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://github.com/MJAHMADEE\" rel=\"noopener noreferrer\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454572134/Github.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://t.me/mjahmadee\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454573000/Telegram.png\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://x.com/MJXAhmadi\" target=\"_blank\"\u003e\u003cimg src=\"https://s29.picofile.com/file/8466445000/X.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://www.facebook.com/mjahmadi.fb\" target=\"_blank\"\u003e\u003cimg src=\"https://s24.picofile.com/file/8454572168/fb.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://www.instagram.com/mjahmadi.ig/\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454572876/insta.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://open.spotify.com/user/z14s6ylxt1nql0us1uny613sd?si=d42deb1a572e4c24\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454572868/Music.png\" alt=\"\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003ca href=\"https://soundcloud.com/mjahmadi\" target=\"_blank\"\u003e\u003cimg src=\"https://s25.picofile.com/file/8454572976/SoundCloud.png\" width=\"25\" height=\"25\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\n---\n\n## 📞 Support \u0026 Community\n\n### Getting Help\n- **📖 Documentation**: Refer to individual README files in each module\n- **🐛 Issues**: [GitHub Issues](https://github.com/MJAHMADEE/Cataract-LMM/issues)\n- **💬 Discussions**: [GitHub Discussions](https://github.com/MJAHMADEE/Cataract-LMM/discussions)\n- **📧 Email**: mjahmadee@gmail.com\n\n---\n\n## 🚀 Roadmap\n\n### Current Version (v1.0.0)\n- ✅ Multi-task surgical video analysis framework\n- ✅ Instance segmentation with YOLO/Mask R-CNN/SAM\n- ✅ Phase recognition with Video Transformers\n- ✅ Skill assessment framework\n- ✅ Production-ready CI/CD pipeline\n\n### Upcoming Features (v1.1.0)\n- 🔄 Real-time inference optimization\n- 🔄 Multi-GPU distributed training\n- 🔄 Model quantization and pruning\n- 🔄 REST API and web interface\n- 🔄 Advanced analytics dashboard\n\n### Future Vision (v2.0.0+)\n- 🔮 Multi-modal learning (video + audio + sensor data)\n- 🔮 Federated learning across institutions\n- 🔮 Real-time surgical guidance system\n- 🔮 Integration with surgical robots\n- 🔮 Multi-language support\n\n---\n\n\u003cdiv align=\"center\"\u003e\n  \u003ch3\u003e🏥 Advancing Surgical AI Through Open Science 🤖\u003c/h3\u003e\n  \u003cp\u003e\n    \u003cstrong\u003eBuilt with ❤️ by the Surgical AI Research Community\u003c/strong\u003e\u003cbr\u003e\n    \u003cem\u003eEmpowering the next generation of computer-assisted surgery\u003c/em\u003e\n  \u003c/p\u003e\n\u003c/div\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmjahmadee%2Fcataract-lmm","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmjahmadee%2Fcataract-lmm","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmjahmadee%2Fcataract-lmm/lists"}