MMM-Face-Reco-DNN receiving PythonShellError: TypeError

KSumwalt · Aug 9, 2020, 5:29 AM

I am trying to get facial recognition working but it is not happening. Here is what I have found. Any advice on how to get this to work?

I have loaded the MMM-Face-Reco-DNN and even used the i_face_recognition.py from https://www.pyimagesearch.com/2018/06/25/raspberry-pi-face-recognition/ with the encodings file generated from the MMM-Face-Reco-DNN instructions and had that .py file recognize faces. So I backed everything out of my config file except the default MMM-Face-Reco-DNN settings from the instructions and when I manually start MM2 everything looks fine. I am running on a Raspberry Pi 4. I did install a different facial recognition module, but it was not in the config file and decided not to use it when I saw it only recognized 1 person. When I step in front of the camera, this comes up in the terminal where I manually started MM2:

[2020-08-09 00:59:28.399] [ERROR]  Whoops! There was an uncaught exception...
[2020-08-09 00:59:28.400] [ERROR]  PythonShellError: TypeError: '<' not supported between instances of 'numpy.ndarray' and 'str'
    at PythonShell.parseError (/home/pi/MagicMirror/modules/MMM-Face-Reco-DNN/node_modules/python-shell/index.js:246:21)
    at terminateIfNeeded (/home/pi/MagicMirror/modules/MMM-Face-Reco-DNN/node_modules/python-shell/index.js:129:32)
    at ChildProcess.<anonymous> (/home/pi/MagicMirror/modules/MMM-Face-Reco-DNN/node_modules/python-shell/index.js:121:13)
    at ChildProcess.emit (events.js:200:13)
    at Process.ChildProcess._handle.onexit (internal/child_process.js:272:12)
    ----- Python Traceback -----
    File "modules/MMM-Face-Reco-DNN/tools/facerecognition.py", line 143, in <module>
      if minDistance < args["tolerance"]: {
  traceback: 'Traceback (most recent call last):\n  File ' +
    '"modules/MMM-Face-Reco-DNN/tools/facerecognition.py", line ' +
    '143, in <module>\n    if minDistance < args["tolerance"]:\n' +
    "TypeError: '< ' not supported between instances of " +
    "'numpy.ndarray' and 'str'\n",
  executable: 'python3',
  options: null,
  script: 'modules/MMM-Face-Reco-DNN/tools/facerecognition.py',
  args: [
    '--cascade=modules/MMM-Face-Reco-DNN/tools/haarcascade_frontalface_default.xml',
    '--encodings=modules/MMM-Face-Reco-DNN/tools/encodings.pickle',
    '--usePiCamera=1',
    '--source=0',
    '--rotateCamera=0',
    '--method=dnn',
    '--detectionMethod=hog',
    '--interval=2000',
    '--output=0',
    '--extendDataset=False',
    '--dataset=modules/MMM-Face-Reco-DNN/dataset/',
    '--tolerance=0.6'
  ],
  exitCode: 1
}
[2020-08-09 00:59:28.409] [ERROR]  MagicMirror will not quit, but it might be a good idea to check why this happened. Maybe no internet connection?
[2020-08-09 00:59:28.409] [ERROR]  If you think this really is an issue, please open an issue on GitHub: https://github.com/MichMich/MagicMirror/issues

Here is my config.js

/* Magic Mirror Config Sample
 *
 * By Michael Teeuw https://michaelteeuw.nl
 * MIT Licensed.
 *
 * For more information on how you can configure this file
 * See https://github.com/MichMich/MagicMirror#configuration
 *
 */

var config = {
	address: "0.0.0.0",
	port: 8080,
	basePath: "/", 	// The URL path where MagicMirror is hosted. If you are using a Reverse proxy
					// you must set the sub path here. basePath must end with a /
	ipWhitelist: [],

	useHttps: false, 		// Support HTTPS or not, default "false" will use HTTP
	httpsPrivateKey: "", 	// HTTPS private key path, only require when useHttps is true
	httpsCertificate: "", 	// HTTPS Certificate path, only require when useHttps is true

	language: "en",
	logLevel: ["INFO", "LOG", "WARN", "ERROR"],
	timeFormat: 12,
	units: "imperial",
	// serverOnly:  true/false/"local" ,
	// local for armv6l processors, default
	//   starts serveronly and then starts chrome browser
	// false, default for all NON-armv6l devices
	// true, force serveronly mode, because you want to.. no UI on this device

	modules: [
		{
			module: "alert",
			classes: 'default everyone',
		},
		{
			module: "updatenotification",
			classes: 'default everyone',
			position: "top_bar"
		},
		{
			module: "clock",
			classes: 'default everyone',
			position: "top_left"
		},
		{
			module: 'MMM-Face-Reco-DNN',
			config: {
			  // Logout 15 seconds after user was not detected any more
			  // If they are detected within this period, the delay will start again
			  logoutDelay: 15000,
			  // How often the recognition starts in milliseconds
			  // With a Raspberry Pi 3+ it works well every 2 seconds
			  checkInterval: 2000,
			  // Module set used for strangers or if no user is detected
			  defaultClass: 'default',
			  // Set of modules which should be shown for every recognised user
			  everyoneClass: 'everyone',
			  // XML to recognize with haarcascade
			  cascade: 'modules/MMM-Face-Reco-DNN/tools/haarcascade_frontalface_default.xml',
			  // Pre-encoded pickle with the faces
			  encodings: 'modules/MMM-Face-Reco-DNN/tools/encodings.pickle',
			  // Use Raspberry Pi camera or another type
			  // 1 = RasPi camera, 0 = other camera
			  usePiCamera: 1,
			  // If using another type of camera, you can choose
			  // i.e. 0 = /dev/video0 or 'http://link.to/live'
			  source: 0,
			  // Rotate camera
			  rotateCamera: 0,
			  // Method of facial recognition
			  // dnn = deep neural network, haar = haarcascade
			  method: 'dnn',
			  // Which face detection model to use
			  // "hog" is less accurate but faster on CPUs
			  // "cnn" is a more accurate deep-learning model which is GPU/CUDA accelerated
			  detectionMethod: 'hog',
			  // How long in milliseconds modules take to hide and show
			  animationSpeed: 0,
			  // Path to Python to run the face recognition
			  // null or '' means default path
			  pythonPath: null,
			  // Should a welcome message be shown using the MagicMirror alerts module?
			  welcomeMessage: true,
			  // Capture new pictures of recognized people, if unknown we save it in folder "unknown"
			  // So you can extend your dataset and retrain it afterwards for better recognitions
			  extendDataset: false,
			  // If extendDataset is true, you need to set the full path of the dataset
			  dataset: 'modules/MMM-Face-Reco-DNN/dataset/',
			  // How much distance between faces to consider it a match. Lower is more strict.
			  tolerance: 0.6
			}
		},


	]
};

/*************** DO NOT EDIT THE LINE BELOW ***************/
if (typeof module !== "undefined") {module.exports = config;}

And here is everything in the terminal before stepping in front of the camera:

pi@raspberrypi:~/MagicMirror $ npm start

> magicmirror@2.12.0 start /home/pi/MagicMirror
> DISPLAY="${DISPLAY:=:0}" ./node_modules/.bin/electron js/electron.js

[2020-08-09 00:58:21.215] [LOG]    Starting MagicMirror: v2.12.0
[2020-08-09 00:58:21.240] [LOG]    Loading config ...
[2020-08-09 00:58:21.248] [LOG]    Loading module helpers ...
[2020-08-09 00:58:21.251] [LOG]    No helper found for module: alert.
[2020-08-09 00:58:21.325] [LOG]    Initializing new module helper ...
[2020-08-09 00:58:21.326] [LOG]    Module helper loaded: updatenotification
[2020-08-09 00:58:21.328] [LOG]    No helper found for module: clock.
[2020-08-09 00:58:21.354] [LOG]    Initializing new module helper ...
[2020-08-09 00:58:21.355] [LOG]    Module helper loaded: MMM-Face-Reco-DNN
[2020-08-09 00:58:21.355] [LOG]    All module helpers loaded.
[2020-08-09 00:58:21.566] [LOG]    Starting server on port 8080 ... 
[2020-08-09 00:58:21.574] [INFO]   You're using a full whitelist configuration to allow for all IPs
[2020-08-09 00:58:21.580] [LOG]    Server started ...
[2020-08-09 00:58:21.581] [LOG]    Connecting socket for: updatenotification
[2020-08-09 00:58:21.583] [LOG]    Connecting socket for: MMM-Face-Reco-DNN
[2020-08-09 00:58:21.583] [LOG]    Starting module helper: MMM-Face-Reco-DNN
[2020-08-09 00:58:21.584] [LOG]    Sockets connected & modules started ...
[2020-08-09 00:58:21.891] [LOG]    Launching application.
[2020-08-09 00:58:25.058] [INFO]   Checking git for module: MMM-Face-Reco-DNN
[2020-08-09 00:58:32.928] [LOG]    [MMM-Face-Reco-DNN] loading encodings + face detector...
[2020-08-09 00:58:33.177] [LOG]    [MMM-Face-Reco-DNN] starting video stream...

sdetweil · Aug 10, 2020, 9:45 PM

@KSumwalt it seems that python has added type checking, which it wasn’t using before.

there was another python related type checking problem a few days ago…

looks like the node_helper has launched the python script to compare the face files, and run into a problem checking the parms passed to it.

KSumwalt · Aug 13, 2020, 5:38 PM

@sdetweil Thank you. I have passed this on as an issue on the github site for the plugin.

x3mEr · Aug 14, 2020, 7:32 PM

@KSumwalt , hi! Check the module update.

KSumwalt · Aug 19, 2020, 2:51 PM

Beautiful! I updated and it appears to be working fine. Thank you, @x3mEr !!!