MagicMirror Forum

    • Register
    • Login
    • Search
    • Recent
    • Tags
    • Unsolved
    • Solved
    • MagicMirror² Repository
    • Documentation
    • Donate
    • Discord

    UNSOLVED MMM-Facial-Recognition-OCV3 -> tools.train.py Problem

    Troubleshooting
    2
    2
    700
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • N
      necRom last edited by

      Hello guys,
      its my first post here. I am not that firm in Python but already had done some things in C. So lets say i´m a beginner in coding and a beginner in Raspberrys.

      I installed MM2 and some modules, which are working after some tweaking. But currently i´m having big problems with the Facial Recognition Module (OCV3).

      “Everything” installed, did it two times on two rasps and got to the same error:
      python tools.capture.py
      does work. I´m capturing images, they are saved and cropped and whatever.

      But when i try
      python train.tools.py
      i get:

      Remember to set the name list environment variable FACE_USERS
      Reading training images...
      --------------------
      Assign label 1 to xxx
      Assign label 2 to xxc
      --------------------
      
      Read 43 positive images and 400 negative images.
      
      7 images from subject xxx
      36 images from subject xxc
      --------------------
      
      Training model with threshold 80
      Traceback (most recent call last):
        File "tools.train.py", line 14, in <module>
          ToolsTrain().train()
        File "/home/pi/MagicMirror/modules/MMM-Facial-Recognition-OCV3/lib/tools/train.py", line 71, in train
          model = ToolsConfig.model()
        File "/home/pi/MagicMirror/modules/MMM-Facial-Recognition-OCV3/lib/tools/config.py", line 99, in model
          return CommonConfig.model(cls.POSITIVE_THRESHOLD)
        File "/home/pi/MagicMirror/modules/MMM-Facial-Recognition-OCV3/lib/common/commonconfig.py", line 51, in model
          model = cv2.face.LBPHFaceRecognizer_create(threshold=thresh)
      AttributeError: 'module' object has no attribute 'face'
      
      

      and with python3 tools.train.py

      Traceback (most recent call last):
        File "tools.train.py", line 10, in <module>
          from lib.tools.train import ToolsTrain
        File "/home/pi/MagicMirror/modules/MMM-Facial-Recognition-OCV3/lib/tools/train.py", line 17, in <module>
          import cv2
        File "/usr/local/lib/python3.5/dist-packages/cv2/__init__.py", line 3, in <module>
          from .cv2 import *
      ImportError: libhdf5_serial.so.100: cannot open shared object file: No such file or directory
      
      

      python2 tools.train.py
      gives the same result as
      python

      I am directly writing on the rasp, not connected by ssh or so.

      opencv, opencv_contrib are installed, pip is there.
      It looks like cv2 is not there … when i use python and then import cv2
      there is no error. But when i then ask for help(cv2.face) the error appears:
      AttributeError: module object has no attribute face

      I got Python 2 (ver 2.7.13) and Python 3 (ver 3.5.3), numpy (ver 1.12.1), OpenCV (ver 3.3.0)

      Maybe i did not install everything needed? But it worked till the train part. Maybe some links are missing or wrong?
      I hope you can bring up some ideas or at best a working solution ^^.

      Thanks
      Nec

      1 Reply Last reply Reply Quote 0
      • B
        Bernard-h last edited by

        Im having the exact same error readings. Frustrating. Has anyone managed to tackle or solve this issue?

        1 Reply Last reply Reply Quote 0
        • 1 / 1
        • First post
          Last post
        Enjoying MagicMirror? Please consider a donation!
        MagicMirror created by Michael Teeuw.
        Forum managed by Paul-Vincent Roll and Rodrigo Ramírez Norambuena.
        This forum is using NodeBB as its core | Contributors
        Contact | Privacy Policy