- Faciem-Deprehendatur is a web application which can do our Real Time AI Face Landmark Detection
- Its a great technique that uses Machine Learning , Deep Learning and Artificial Intelligance to detect different points on our face
- The points detected when combined together generates a mesh on the face . The Face Detector in this project detects 486 3D facial landmarks to infer the approximate surface geometry of a human face .
Now its your turn to play 🤹♂️ here 👉 faciem-deprehendatur
The project requires usage of pre-trained model by TensorflowJs named as " FaceMesh 😊 " , it simply detects our face in real time and shows a 3D mesh on our face
- A webcam stream is created inside a front-end React App
- FaceMesh model is loaded from TensorFlow.Js Library and detections are made from the webcam stream running
- Keypoints and Triangles are drawn on JavaScript Canvas [The Blue dots show the 3D Points and the Black Lines are the Neural Networks joining those points ]
| Example 1 | Example 2 | Example 3 |
|---|---|---|
 |
 |
 |
The project can be coupled with Iris Detection System to track movements of Iris with the face as well which can used to predict the actions of the targeted person by evaluating his/her expressions.
-
Clone / Download :
git clone https://github.com/your_username_/Project-Name.git -
Open the project in your favorite code editor and do following
npm install @tensorflow/tfjs @tensorflow-models/face-landmarks-detection react-webcam
Contributions are what make the open source community such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated.
- Fork the Project
- Create your new Branch
git checkout -b NewBranchName - Commit your Changes
git commit -m " Description of contribution you made " - Push to the Branch
git push origin NewBranchName - Open a Pull Request