houloude9's picture
Update README.md
432aaa6 verified
---
title: FaceRecognity
emoji: 🐢
colorFrom: pink
colorTo: green
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false
python_version: "3.10"
short_description: Facial recognition with InsightFace using Gradio
tags: ["computer-vision", "facial-recognition", "insightface", "gradio"]
fullWidth: true
---
# FaceRecognity 🐢
FaceRecognity is a **facial recognition web app** built with **InsightFace** and **Gradio**, designed to run smoothly on Hugging Face Spaces.
It allows you to:
- **Extract face embeddings** from any uploaded image.
- **Compare a target face** against multiple candidate images.
- Get **match confidence scores** for similarity.
---
## Features
1. **Extract Embeddings**
- Upload an image and get a numerical face embedding vector.
- Useful for storing face representations or running offline comparisons.
2. **Match Faces**
- Upload a target face image and multiple candidate images.
- Quickly identify which candidates match the target face.
- Returns similarity scores (0–100%).
---
## How to Use
1. **Extract Embedding**
- Go to the "Extract Embedding" tab.
- Upload an image with a clear face.
- Click **Extract** to get the embedding.
2. **Match Faces**
- Go to the "Match Faces" tab.
- Upload a target image and up to 5 candidate images.
- Click **Match** to see which candidates match the target.
---
## Requirements
This Space uses the following Python packages:
- `gradio`
- `numpy`
- `opencv-python-headless`
- `insightface`
- `torch`
- `onnxruntime`
- `tensorflow-cpu`
> The `requirements.txt` is already included in the repository for automatic installation.
---
## Notes
- Works best with **single-face images** for extraction and matching.
- Candidate images without a detected face are ignored.
- Runs entirely in-memory—no images are stored on the server.
---
## References
- [InsightFace GitHub](https://github.com/deepinsight/insightface)
- [Gradio Docs](https://gradio.app)
- [Hugging Face Spaces Configuration Reference](https://huggingface.co/docs/hub/spaces-config-reference)