{"id":23706967,"url":"https://github.com/kr1shnasomani/DeepDetect","last_synced_at":"2025-09-03T12:32:19.388Z","repository":{"id":269799610,"uuid":"908496578","full_name":"kr1shnasomani/SkySight","owner":"kr1shnasomani","description":"Traffic light, vehicle and human detection from aerial images using YOLOv8 model and Computer Vision","archived":false,"fork":false,"pushed_at":"2024-12-26T08:12:03.000Z","size":1096,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-12-26T09:19:40.070Z","etag":null,"topics":["computer-vision","deep-learning","neural-network","numpy","opencv","yolov8"],"latest_commit_sha":null,"homepage":"","language":"Python","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/kr1shnasomani.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":"2024-12-26T08:04:04.000Z","updated_at":"2024-12-26T08:16:09.000Z","dependencies_parsed_at":"2024-12-26T09:19:43.068Z","dependency_job_id":"4505bd2e-39be-4d82-8954-5eaf6295ac37","html_url":"https://github.com/kr1shnasomani/SkySight","commit_stats":null,"previous_names":["kr1shnasomani/skysight"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kr1shnasomani%2FSkySight","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kr1shnasomani%2FSkySight/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kr1shnasomani%2FSkySight/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kr1shnasomani%2FSkySight/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kr1shnasomani","download_url":"https://codeload.github.com/kr1shnasomani/SkySight/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":231882156,"owners_count":18440325,"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","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":["computer-vision","deep-learning","neural-network","numpy","opencv","yolov8"],"created_at":"2024-12-30T16:01:35.557Z","updated_at":"2025-09-03T12:32:19.364Z","avatar_url":"https://github.com/kr1shnasomani.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003ch1 align=\"center\"\u003eDeepDetect\u003c/h1\u003e\nThe deepfake detection system extracts frames from videos, detects and crops faces using facenet-pytorch, and classifies them as real or fake using a deep learning model like EfficientNet. The system highlights deepfake faces and provides visual analysis, leveraging PyTorch, OpenCV, and timm for model implementation.\n\n## Execution Guide:\n1. Run the following command line in the terminal:\n   ```\n   pip install torch torchvision torchaudio facenet-pytorch timm opencv-python numpy matplotlib pillow\n   ```\n\n2. Enter the path of the video\n\n3. Upon running all the cells the code will provide its prediction frame by frame\n\n## Result:\n\n  DeepFake Video:\n  \n  [deepfake](https://github.com/user-attachments/assets/aa03466e-0030-4d5e-a640-0094ffb00c26)\n\n  `Model prediction: DeepFake`\n\n  Original Video:\n  \n  [original](https://github.com/user-attachments/assets/0304a40d-954e-4dfc-b0a9-5ccabe49e69a)\n\n  `Model prediction: Original`\n\n## Overview:\nThe above code performs the following key tasks:  \n\n1. **Face Detection**:  \n   Uses **MTCNN** (Multi-task Cascaded Convolutional Networks) from `facenet_pytorch` to detect faces in images or video frames.  \n\n2. **Deepfake Classification Model**:  \n   - Loads a **pretrained Xception model** from the `timm` library, which is widely used for deepfake detection.  \n   - The model is fine-tuned with a single output neuron (binary classification: real or fake).  \n\n3. **Preprocessing**:  \n   - Detected faces are extracted and preprocessed before being passed to the model for classification.  \n\n4. **Inference**:  \n   The model predicts whether a given face is real or deepfake.  \n\nThis approach leverages **CNN-based face detection** and **deep learning classification** to identify manipulated media.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkr1shnasomani%2FDeepDetect","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkr1shnasomani%2FDeepDetect","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkr1shnasomani%2FDeepDetect/lists"}