{"id":29916281,"url":"https://github.com/amdjadouxx/bones_break_detection","last_synced_at":"2025-08-02T04:43:33.965Z","repository":{"id":306011104,"uuid":"1023623181","full_name":"amdjadouxx/BONES_BREAK_DETECTION","owner":"amdjadouxx","description":"usage of yolo to detect if bones are break or not","archived":false,"fork":false,"pushed_at":"2025-07-23T06:30:23.000Z","size":5875,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-07-23T07:20:05.006Z","etag":null,"topics":["artificial-intelligence","image-processing","yolov8"],"latest_commit_sha":null,"homepage":"","language":"Python","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/amdjadouxx.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-07-21T12:46:46.000Z","updated_at":"2025-07-23T06:30:27.000Z","dependencies_parsed_at":"2025-07-23T07:20:13.781Z","dependency_job_id":"b8d41816-996e-4f7d-a8d9-feb814a5e763","html_url":"https://github.com/amdjadouxx/BONES_BREAK_DETECTION","commit_stats":null,"previous_names":["amdjadouxx/bones_break_detection"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/amdjadouxx/BONES_BREAK_DETECTION","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/amdjadouxx%2FBONES_BREAK_DETECTION","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/amdjadouxx%2FBONES_BREAK_DETECTION/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/amdjadouxx%2FBONES_BREAK_DETECTION/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/amdjadouxx%2FBONES_BREAK_DETECTION/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/amdjadouxx","download_url":"https://codeload.github.com/amdjadouxx/BONES_BREAK_DETECTION/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/amdjadouxx%2FBONES_BREAK_DETECTION/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":268334834,"owners_count":24233833,"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":"2025-08-02T02:00:12.353Z","response_time":74,"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":["artificial-intelligence","image-processing","yolov8"],"created_at":"2025-08-02T04:43:26.778Z","updated_at":"2025-08-02T04:43:33.943Z","avatar_url":"https://github.com/amdjadouxx.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🦴 Pediatric Bone Fracture Detection with YOLOv8\n\n[![Python 3.8+](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)\n[![YOLOv8](https://img.shields.io/badge/YOLOv8-Ultralytics-orange.svg)](https://github.com/ultralytics/ultralytics)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n\n🏥 **Détection automatique de fractures osseuses chez les enfants** à partir de radiographies médicales utilisant YOLOv8.\n\n## 📋 Table des matières\n- [🎯 Objectif](#-objectif)\n- [🚀 Installation rapide](#-installation-rapide)\n- [📁 Structure du projet](#-structure-du-projet)\n- [🔧 Utilisation](#-utilisation)\n- [📊 Dataset](#-dataset)\n- [🧠 Modèle](#-modèle)\n- [🖼️ Interface Web](#-interface-web)\n- [📈 Évaluation](#-évaluation)\n- [🤝 Contribution](#-contribution)\n\n## 🎯 Objectif\n\nCe projet propose une **pipeline complète d'inférence** pour détecter automatiquement les fractures visibles sur les radiographies pédiatriques du poignet et de l'avant-bras. \n\n### Fonctionnalités principales :\n- ✅ Détection de fractures avec YOLOv8\n- ✅ Interface web interactive (Streamlit)\n- ✅ Support image/vidéo/webcam\n- ✅ Métriques d'évaluation complètes\n- ✅ Pipeline prête pour la production\n\n## 🚀 Installation rapide\n\n### Prérequis\n- Python 3.8+ \n- GPU recommandé (CUDA compatible)\n\n### Installation\n```bash\n# Cloner le repository\ngit clone https://github.com/amdjadouxx/BONES_BREAK_DETECTION.git\ncd BONES_BREAK_DETECTION\n\n# Installer les dépendances\npip install -r requirements.txt\n\n# Télécharger le dataset (optionnel)\npython scripts/prepare_data.py --download\n\n# Lancer l'interface web\nstreamlit run webapp/app.py\n```\n\n## 📁 Structure du projet\n\n```\n📦 BONES_BREAK_DETECTION/\n├── 📁 data/                    # Gestion du dataset\n│   ├── raw/                    # Données brutes\n│   ├── processed/              # Données prétraitées\n│   └── annotations/            # Annotations YOLO\n├── 📁 models/                  # Modèles YOLOv8\n│   ├── pretrained/             # Modèles pré-entraînés\n│   └── checkpoints/            # Checkpoints d'entraînement\n├── 📁 inference/               # Pipeline de prédiction\n│   ├── predict.py              # Prédiction sur images\n│   ├── video_inference.py      # Prédiction vidéo/webcam\n│   └── batch_predict.py        # Prédictions par lots\n├── 📁 webapp/                  # Interface Streamlit\n│   ├── app.py                  # Application principale\n│   └── components/             # Composants UI\n├── 📁 notebooks/               # Analyse exploratoire\n│   ├── 01_data_exploration.ipynb\n│   ├── 02_model_evaluation.ipynb\n│   └── 03_results_analysis.ipynb\n├── 📁 utils/                   # Fonctions utilitaires\n│   ├── data_utils.py           # Prétraitement données\n│   ├── model_utils.py          # Gestion modèles\n│   ├── visualization.py        # Visualisations\n│   └── metrics.py              # Métriques d'évaluation\n├── 📁 scripts/                 # Scripts utilitaires\n│   ├── prepare_data.py         # Préparation dataset\n│   ├── convert_to_yolo.py      # Conversion format YOLO\n│   └── evaluate_model.py       # Évaluation modèle\n├── 📁 config/                  # Configurations\n│   ├── model_config.yaml       # Config modèle\n│   └── data_config.yaml        # Config dataset\n├── 📄 requirements.txt         # Dépendances Python\n├── 📄 setup.py                 # Installation package\n└── 📄 README.md                # Documentation\n```\n\n## 🔧 Utilisation\n\n### Prédiction sur une image\n```python\nfrom inference.predict import PediatricFractureDetector\n\n# Initialiser le détecteur\ndetector = PediatricFractureDetector('models/best.pt')\n\n# Prédiction\nresults = detector.predict('path/to/xray.jpg')\ndetector.save_results(results, 'output/')\n```\n\n### Prédiction via ligne de commande\n```bash\n# Image unique\npython inference/predict.py --source path/to/image.jpg --output results/\n\n# Dossier d'images\npython inference/predict.py --source path/to/images/ --output results/\n\n# Vidéo ou webcam\npython inference/video_inference.py --source webcam --output results/\n```\n\n## 📊 Dataset\n\nCe projet utilise le **RSNA Pediatric Bone Age Dataset** et des datasets open-source de radiographies pédiatriques :\n\n### Sources de données :\n- 🔗 [RSNA Bone Age Challenge](https://www.rsna.org/education/ai-resources-and-training/ai-image-challenge/rsna-pediatric-bone-age-challenge-2017)\n\n### Caractéristiques :\n- **~12,000 radiographies** pédiatriques\n- **Annotations** : bounding boxes des fractures\n- **Métadonnées** : âge, sexe, localisation anatomique\n- **Format** : DICOM → PNG/JPG + annotations YOLO\n\n## 🧠 Modèle\n\n### Architecture YOLOv8\n- **Modèle base** : YOLOv8n/s/m/l/x (configurable)\n- **Classes** : `fracture`, `no_fracture`\n- **Input** : 640x640 pixels\n- **Pré-traitement** : normalisation, augmentation\n\n### Performance attendue :\n- **Précision** : ~85-90%\n- **Rappel** : ~80-85%\n- **F1-Score** : ~82-87%\n- **Temps d'inférence** : \u003c100ms (GPU)\n\n## 🖼️ Interface Web\n\nInterface Streamlit intuitive permettant :\n- 📤 Upload d'images médicales\n- 🔍 Visualisation des prédictions\n- 📊 Affichage des scores de confiance\n- 💾 Téléchargement des résultats\n- 📈 Historique des prédictions\n\n## 📈 Évaluation\n\n### Métriques disponibles :\n- Précision, Rappel, F1-Score\n- Matrice de confusion\n- Courbes ROC/PR\n- Heatmaps de confiance\n- Analyse des erreurs\n\n### Lancer l'évaluation :\n```bash\npython scripts/evaluate_model.py --model models/best.pt --data data/test/\n```\n\n### Développement local :\n```bash\n# Créer un environnement virtuel\npython -m venv venv\nsource venv/bin/activate  # Windows: venv\\Scripts\\activate\n\n# Installation en mode développement\npip install -e .\n\n# Tests\npython -m pytest tests/\n```\n\n---\n\n## 📞 Contact\n\n- **Auteur** : Amdjadouxx\n- **Email** : amdjadahmodali974@gmail.com\n\n## 📜 License\n\nCe projet est sous licence MIT. Voir [LICENSE](LICENSE) pour plus de détails.\n\n---\n\n⭐ **N'hésitez pas à starred ce projet si vous le trouvez utile !**\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Famdjadouxx%2Fbones_break_detection","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Famdjadouxx%2Fbones_break_detection","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Famdjadouxx%2Fbones_break_detection/lists"}