face_recognition

Recognizes the faces detected by face_detection with the face recognizers of opencv_contrib (Eigenfaces, Fisherfaces or LBPH), trained with the ~/train service.

opencv_apps	1.12.0 and later (ROS 1), 3.0.0 and later (ROS 2)
ROS	noetic, ROS One, jazzy, kilted, lyrical and later
OpenCV	any version of the distributions
Group	Faces, the nodes of which publish ~/faces (FaceArrayStamped)

Usage

# ROS 2
ros2 launch opencv_apps face_recognition.launch.xml image:=/camera/image_raw debug_view:=false
# ROS 1
roslaunch opencv_apps face_recognition.launch image:=/camera/image_raw debug_view:=false use_opencv4:=true

Try it with the demo video

Download the input video of the demo above, publish it as image with image_publisher, and run the node on it, in the terminals:

wget https://raw.githubusercontent.com/ros-perception/opencv_apps/indigo/test/face_detector_withface_test_diamondback.mp4
wget https://raw.githubusercontent.com/ros-perception/opencv_apps/indigo/test/face_data.tar.gz
tar xzf face_data.tar.gz
# terminal 1: publish the video as image in a loop
ros2 run image_publisher image_publisher_node face_detector_withface_test_diamondback.mp4 --ros-args -p publish_rate:=10.0 -r image_raw:=image
# terminal 2: run the node with the parameters of the demo
ros2 launch opencv_apps face_recognition.launch.xml image:=image data_dir:=$PWD/face_data launch_face_detection:=true launch_trainer:=false
# terminal 3: view the result
ros2 run rqt_image_view rqt_image_view /face_recognition/debug_image

The faces trained in test/face_data.tar.gz; ~debug_image is shown.

See Credits of the demo videos for the license of the video.

Subscribed topics

Topic	Type
faces	opencv_apps/msg/FaceArrayStamped
image	sensor_msgs/msg/Image

Published topics

Topic	Type	Note
~/debug_image	sensor_msgs/msg/Image
~/faces (recommended)	opencv_apps/msg/FaceArrayStamped	Recommended, opencv_apps 3.0.1 and later. The same as ~/output, with the name of the other face nodes.
~/output (compatibility)	opencv_apps/msg/FaceArrayStamped	Kept for compatibility, use ~/faces.

Services

Service	Type
~/train	opencv_apps/srv/FaceRecognitionTrain

Parameters

See also the common parameters.

Name	Type	Default (range)	Changed at run time	Description
approximate_sync	bool	false	no, restart the node
data_dir	string	“~/.ros/opencv_apps/face_data”	yes	Save directory for train data
face_model_height	int	90 (30 - 500)	yes	Height of training face image
face_model_width	int	190 (30 - 500)	yes	Width of training face image
face_padding	double	0.1 (0 - 2)	yes	Padding ratio of each face
lbph_grid_x	int	8 (1 - 30)	yes	grid_x parameter used only for LBPH model
lbph_grid_y	int	8 (1 - 30)	yes	grid_y parameter used only for LBPH model
lbph_neighbors	int	8 (1 - 30)	yes	Neighbors parameter used only for LBPH model
lbph_radius	int	1 (1 - 10)	yes	Radius parameter used only for LBPH model
model_method	string	“eigen”	yes	Method to recognize faces, “eigen”: eigen, “fisher”: fisher, “LBPH”: LBPH
model_num_components	int	0 (0 - 100)	yes	Number of components for face recognizer model
model_threshold	double	8000.0 (0 - 10000)	yes	Threshold for face recognizer model
queue_size	int	100	no, restart the node	Specigy queue_size of input image subscribers
save_train_data	bool	true	yes	Save train data
use_saved_data	bool	true	yes	Use saved data

Source

src/nodelet/face_recognition_nodelet.cpp, launch/face_recognition.launch.xml