Spaces:
Sleeping
Sleeping
title: PrimateFace | |
emoji: 🐵ྀི | |
colorFrom: purple | |
colorTo: red | |
sdk: gradio | |
sdk_version: 5.35.0 | |
app_file: app.py | |
pinned: true | |
license: mit | |
tags: | |
- computer-vision | |
- object-detection | |
- pose-estimation | |
- gaze-tracking | |
- primates | |
- animals | |
short_description: Test cross-primate face detection and facial landmark models | |
# PrimateFace Demo | |
State-of-the-art primate face detection, pose estimation, and gaze tracking system. | |
## Features | |
- 🎯 **Multi-species face detection** - Trained on diverse primate species | |
- 📍 **68-point facial landmarks** - Detailed facial structure analysis | |
- 👀 **Gaze estimation** - Experimental gaze direction tracking | |
- 🖼️ **Image & Video support** - Process both static images and videos | |
- 📸 **Webcam capture** - Real-time processing from webcam | |
## Model Architecture | |
- **Detection Backend**: MMDetection with Cascade R-CNN (ResNet101-FPN) | |
- **Pose Estimation**: MMPose with HRNet-W18 (68 keypoints) | |
- **Gaze Estimation**: Gazelle model (trained only on human faces) | |
## Usage | |
1. Upload an image/video or capture from webcam | |
2. Select the task type (detection only, with pose, or with gaze) | |
3. Adjust confidence threshold if needed | |
4. Click "Detect Faces" to process | |
## Note | |
This demo runs on a dedicated GPU server. If the server is offline, please check back later. | |