Facial Recognition - customize your mirror for every user!

PointPubMedia

@paviro Do you think it can be possible to have a “CURL” call when a user log in and log out ?

bhepler

@YourhighnessLeender Check your capitalization. Your training script output was ‘Johannes’ and your config.js has ‘johannes’.

YourhighnessLeender

@bhepler thanks! Tried that but still no luck. Started over from scratch, still undefined user. The. Lowercase was me testing if it made any difference.

Cound it be that the webcam captures in too low res when running MM compare to the test script? And in that case, how can I check or change that?

Im a noob with linux and the terminal.

nuckinfutz

@YourhighnessLeender You spelled your name as Johannes in the training file and in config.js you spelled it johannes. config.js is a lower case J.

Been there done that. lol…

nuckinfutz

@nuckinfutz behelper beat me to it.

benkej

Hi guys,

I have the same problem, tried with picam and normal webcam too.
Capture working, train working, facerecognition_tester working, MMM-Facial-Recognition doesn’t work.

Anyone sitting in front of the mirror, I get the message in terminal “User undefined with confidence null logged in”

nuckinfutz

@YourhighnessLeender Webcam is fine. If it worked in training it will work in the module.

YourhighnessLeender

@nuckinfutz that is whats so strange. It does work in training but not in the module. Did a new install of the module and new training.xml but no luck. Very very wierd.

Maybe I should buy a picam and try with that.

YourhighnessLeender

@benkej question: did you change any of the capture settings in config.js? Like threshold etc? I was thinking of trying a different algorithm, did you try that?

benkej

@YourhighnessLeender only for trial but I changed everything back to the default. I’m using raspberry pi 3 model b, here is part of my config:

{
    module: 'MMM-Facial-Recognition',
    config: {
        // 1=LBPH | 2=Fisher | 3=Eigen
        recognitionAlgorithm: 1,
        // Threshold for the confidence of a recognized face before it's considered a
        // positive match.  Confidence values below this threshold will be considered
        // a positive match because the lower the confidence value, or distance, the
        // more confident the algorithm is that the face was correctly detected.
        lbphThreshold: 50,
        fisherThreshold: 250,
        eigenThreshold: 3000,
        // force the use of a usb webcam on raspberry pi (on other platforms this is always true automatically)
        useUSBCam: true,
        // Path to your training xml
        trainingFile: 'modules/MMM-Facial-Recognition/training.xml',
        // recognition intervall in seconds (smaller number = faster but CPU intens!)
        interval: 3,
        // Logout delay after last recognition so that a user does not get instantly logged out if he turns away from the mirror for a few seconds
        logoutDelay: 15,
        // Array with usernames (copy and paste from training script)
        users: ["Jani"],
        //Module set used for strangers and if no user is detected
        defaultClass: "default",
        //Set of modules which should be shown for every user
        everyoneClass: "everyone"
    }
}

and part of console log:

[MMM-Facial-Recognition] Facerecognition started...
[MMM-Facial-Recognition] Loading training data...
[MMM-Facial-Recognition] ALGORITHM: LBPH
[MMM-Facial-Recognition] Training data loaded!
[MMM-Facial-Recognition] --------------------
[MMM-Facial-Recognition] Webcam ausgewählt...
[MMM-Facial-Recognition] User undefined with confidence null logged in.
[MMM-Facial-Recognition] User undefined logged out.
[MMM-Facial-Recognition] User undefined with confidence null logged in.
[MMM-Facial-Recognition] User undefined logged out.
[MMM-Facial-Recognition] User undefined with confidence null logged in.
[MMM-Facial-Recognition] User undefined logged out.
^C[MMM-Facial-Recognition] Shutdown: Cleaning up camera...
[MMM-Facial-Recognition] Terminating..

@YourhighnessLeender and the output of train.py:

Which algorithm do you want to use?
[1] LBPHF (recommended)
[2] Fisherfaces
[3] Eigenfaces
--> 1

Reading training images...
--------------------
Assign label 1 to Jani
--------------------

Read 3 positive images and 400 negative images.

3 images from subject Jani
--------------------

Training model...
Training data saved to training.xml

Please add or update (if you added new people not just new images) ['Jani'] inside config.js (mirror module) or config.py (model tester). You can change the names to whatever you want, just keep the same order and you'll be fine.
Please add 1 as your choosen algorithm inside config.js (mirror module) or config.py (model tester).

Note from admin: Please use Markdown on code snippets for easier reading!

YourhighnessLeender

@benkej same as me, but no luck :-(

benkej

@YourhighnessLeender unfortunately yes :angry:
I try to reinstall the whole system

YourhighnessLeender

@benkej oke, please update on your progress. Good luck. Im very close to giving up.

paviro

@PointPubMedia Well sure but you could also write your own module listening for the notifications from my module :)

@bhepler said in Facial Recognition - customize your mirror for every user!:

@YourhighnessLeender Check your capitalization. Your training script output was ‘Johannes’ and your config.js has ‘johannes’.

This does not make any difference :)

@benkej said in Facial Recognition - customize your mirror for every user!:

@YourhighnessLeender only for trial but I changed everything back to the default. I’m using raspberry pi 3 model b, here is part of my config:
Read 3 positive images and 400 negative images.

You should definitely use more than 3 pictures! I use around 200 :D

YourhighnessLeender

@paviro 200 pictures - aha! Maybe my 13 isnt enough. I can’t use the capture script since i’m on webcam so I took em with my phone and transferred them to the pi. I’ll add more and cross my fingers.

paviro

Well that could also be a cause. You should use the same camera for best results!
Why shouldn’t you be able to use a webcam with the capture script? I never used it with anything else to be honest :D

YourhighnessLeender

@paviro hmm. Just can’t get it working for some reason… will try again. It complaind about no picam. Maybe I have to change something in the script? I’ll du another try.

paviro

Well the capture script does not have an enforce webcam option. Try to uninstall the picam python module, it will fallback to webcam then.

To minimize recognition errors the MagicMirror module requires two identical recognitions to be accepted as current user. If your training data is not good enough it could happen that you are never recognized twice in two directly following images -> Always unknown user.

YourhighnessLeender

@paviro thanks for the help. How do I uninstall the picam? Is it just to remove picam.py?

paviro

Nooo! Don’t mess with the module code! I just added an option to force webcam use, checkout the updated code on GitHub :)