mrlathome / faceutils Goto Github PK

Hi,

when we are tracking the SAME face if people changed is the tracker still tracking face (different people)

How we can detect people and scene changed ?

Best

Hi,
I've copied and renamed 'ros' folder inside a ros workspace.
Build was ok, but when I run the launch file the following error appears:
process[ros_face_recognition-2]: started with pid [9948] Traceback (most recent call last): File "/home/luca/ros/src/ros_face_recognition/scripts/face_recognition_node.py", line 2, in <module> import cv2 ImportError: /opt/ros/kinetic/lib/python2.7/dist-packages/cv2.so: undefined symbol: PyCObject_Type

This is my configuration
Ubuntu: 16.04
ROS: Kinetic
OpenCV 3.4.6
Python 2.7

All my cpp code run without problems, but I'm not able to integrate python code importng cv2 libraries insied ROS environment.

When I run face_recognition_node.py , now I receive this message and the node dies.

process[ros_face_recognition-1]: started with pid [14644]
Traceback (most recent call last):
  File "/home/luca/ros/src/ros_face_recognition/scripts/face_recognition_no
    import face_api
  File "/home/luca/ros/src/ros_face_recognition/scripts/face_api.py", line
    _classifier = pickle.load(open(os.path.join(os.path.join(_base_dir, "da
TypeError: a bytes-like object is required, not 'str'

May be the file gender_model.pickle should be in binay format instead of text ?

After running the face recognition node once, and attempting to launch it again, Python crashes (no log or error message)

By commenting the scripts, I managed to find out that the crash happens when pickle tries to load the .dump files generated from the previous session.

I also found out that those errors are related with the loading and saving of c pointers of python-wrapped c classes. (I suspect that the newer version of dlib I'm using is the culprit)

I'm using dlib 19.10, if you could tell me your version, I'd try to use it and validate my hypothesis.

Thanks!

hi,
unable to import rospy
ImportError: No module named 'rospy'

please help,
thanks and regards.