{"id":28306655,"url":"https://github.com/harshramani00/human-action-recognition","last_synced_at":"2025-07-24T22:40:55.772Z","repository":{"id":281203639,"uuid":"944540452","full_name":"harshramani00/Human-Action-Recognition","owner":"harshramani00","description":"A Human Action Recognition (HAR) model combining 3D CNN and LSTM networks to accurately recognize actions in videos using spatial-temporal feature extraction. Trained on UCF-50 and outperforming existing architectures.","archived":false,"fork":false,"pushed_at":"2025-03-07T14:36:56.000Z","size":3830,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"main","last_synced_at":"2025-06-20T04:38:38.817Z","etag":null,"topics":["3d-cnn","computer-vision","deep-learning","human-action-recognition","lstm","machine-learning","python","spatial-temporal","tensorflow","ucf50-dataset","video-classification"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/harshramani00.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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}},"created_at":"2025-03-07T14:26:49.000Z","updated_at":"2025-05-20T02:05:26.000Z","dependencies_parsed_at":"2025-03-07T15:43:28.730Z","dependency_job_id":null,"html_url":"https://github.com/harshramani00/Human-Action-Recognition","commit_stats":null,"previous_names":["harshramani00/human-action-recognition"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/harshramani00/Human-Action-Recognition","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/harshramani00%2FHuman-Action-Recognition","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/harshramani00%2FHuman-Action-Recognition/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/harshramani00%2FHuman-Action-Recognition/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/harshramani00%2FHuman-Action-Recognition/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/harshramani00","download_url":"https://codeload.github.com/harshramani00/Human-Action-Recognition/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/harshramani00%2FHuman-Action-Recognition/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":266914588,"owners_count":24005594,"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":"2025-07-24T02:00:09.469Z","response_time":99,"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":["3d-cnn","computer-vision","deep-learning","human-action-recognition","lstm","machine-learning","python","spatial-temporal","tensorflow","ucf50-dataset","video-classification"],"created_at":"2025-05-24T04:12:23.457Z","updated_at":"2025-07-24T22:40:55.764Z","avatar_url":"https://github.com/harshramani00.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Swift-Spatio Flow: A Human Action Recognition Model Using 3D CNN-LSTM\n\n## 📌 Project Overview\n\n**Swift-Spatio Flow** is an advanced **Human Action Recognition (HAR)** model that combines **3D Convolutional Neural Networks (3D CNN)** with **Long Short-Term Memory (LSTM)** networks. This project aims to improve action recognition in videos by efficiently extracting **spatial and temporal features** while reducing computational cost.\n\n🚀 **Key Applications**\n- CCTV surveillance enhancement\n- Assisting the visually impaired\n- Self-driving cars\n- Sports analytics\n\n## 🎯 Problem Statement\nExisting HAR models suffer from:\n- **Complexity:** High computational cost\n- **Accuracy:** Difficulty in handling low-quality videos\n- **Scalability:** Struggle with large datasets\n\n**Swift-Spatio Flow** addresses these challenges by integrating a **3D CNN and LSTM** to extract spatial and temporal features efficiently.\n\n## 📊 Methodology\n1. **Preprocessing:** \n   - Extract frames from videos\n   - Resize and normalize images\n   - Convert frames into sequences\n\n2. **Model Architecture:** \n   - 3D CNN for feature extraction\n   - LSTM for sequence modeling\n   - Softmax activation for classification\n\n3. **Training \u0026 Evaluation:**\n   - Dataset: **UCF-50**\n   - Metrics: **Accuracy, Precision, Recall, F1-score**\n   - Comparison with existing models\n\n## 🏆 Results\n| Model                 | Accuracy (%) | Precision (%) | Recall (%) | F1 Score (%) |\n|----------------------|-------------|-------------|-----------|-------------|\n| CNN + LSTM          | 76.12       | 75.94       | 74.17     | 75.86       |\n| ConvLSTM2D         | 78.95       | 78.74       | 76.14     | 78.68       |\n| Time Distributed CNN | 88.50       | 88.00       | 87.52     | 87.71       |\n| 3D CNN (UCF-101)    | 91.65       | 89.96       | 90.82     | 91.10       |\n| **Swift-Spatio Flow** | **94.89**  | **94.37**  | **93.45** | **93.56** |\n\n## 🔮 Future Enhancements\nTrain on larger datasets like Kinetics for better generalization\nOptimize computational cost for real-time performance\nDeploy the model as a web application\n\n\n## 🤝 Contributors\n\n- Ian Joseph K\n- Aryan Patil (https://github.com/aryanator)\n- Abhishek Raje\n- Ramani Harsh Anilkumar\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fharshramani00%2Fhuman-action-recognition","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fharshramani00%2Fhuman-action-recognition","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fharshramani00%2Fhuman-action-recognition/lists"}