ROS2 package for using face_recognition.
- ROS 2 (Tested on Humble, would work on any distributions with Python 3.3+ required by
face_recognition)
Other dependencies are listed in package.xml and requirements.txt.
$ cd ~/ros2_ws/src
$ git clone https://github.com/mitukou1109/face_recognition_ros.git
$ cd face_recognition_ros
$ rosdep install -iyr --from-paths .
$ pip install -r requirements.txt
$ cd ~/ros2_ws
$ colcon build --symlink-install
$ source install/local_setup.bash
Prepare an image containing a face to identify. Assuming you have your selfie in home directory:
$ cd ~/ros2_ws/src/face_recognition_ros
$ python register_known_face.py ~/your_name.jpg "Your name" # You can specify the name to register, otherwise the file's name will be used
Registered Your name
You can test with the following commands if you have a webcam connected to your PC:
$ sudo apt install ros-humble-v4l2-camera
$ ros2 launch face_recogntion_ros demo.launch.py
...
[INFO] [1703600037.356322923] [face_identifier]: Registered known faces: ['Your name']
...
TODO
Subscribed topics:
| Name | Type | Description |
|---|---|---|
/image_raw |
sensor_msgs/Image |
Input image |
Published topics:
| Name | Type | Description |
|---|---|---|
~/detections |
vision_msgs/Detection2DArray |
Sequence of bounding box and name for each detected faces |
~/result_image |
sensor_msgs/Image |
Input image with bounding boxes and names drawn |
| Name | Type | Default value | Description |
|---|---|---|---|
resize_height |
int | 320 |
Height to resize input image before encoding for making it faster |
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent