I’m running on an RPi 3b and have 4 users. Each user has just under 100 pictures that were used to train the mirror. Giving a total of about 350 images. I have my own module that if the user is recognized it plays a video. (Classic Disney “Mirror Mirror on the wall” type module)
How many images are you using for your recognition file?
@lepetitnicolas Just saw this but here is the fix
cd home/pi/MagicMirror/modules/MMM-Facial-Recognition-Tools
sudo pip install future --upgrade
@Chiumanfu autostart is a folder. You want to drop the mm.desktop file into that folder. Haven’t tried Jessie light but don’t know why it shouldnt work.
It should have worked where you tried as well. POST the contents of your mm.desktop file
I’m not sure what the benefit of PM2 is. In my mind if the mirror crashed the error will need to be fixed and the Pi rebooted. In the event of a power outage the Pi will reboot on its own. This is a simpler way to autostart MM2
In my case I’m running a 7" RPi touchscreen display connected to a RPi v3 running Raspbian Jessie via GPIO pins and the HDMI connector. I am supplying power to both the display and the RPi via a single 2.5v USB power supply. For me, I need to be aware of power consumption.
When I first started I installed PM2 and after some headaches I got it running. It seemed like overkill in my mind even without my need to watch out for power hungry daemons. I figured there must be an easier way. After a lot of searching and trying different things here is what I came up with. It’s very simple and doesn’t carry any of the overhead that PM2 does.
Here’s my solution
Step 1.
Create a text file that’s called mm.sh in the MagicMirror folder. Add the following lines to that file
cd /home/pi/MagicMirror
DISPLAY=:0 npm start
Now save the file.
Right click on it and select Properties. Make sure its set as “executable by anyone” Now close the properties panel.
Step 2.
Create an empty file on your desktop call mm.desktop
open the file with your text editor and copy the following into it
[Desktop Entry]
Version=x.y
Name=MagicMirror
Comment=This is my comment
Exec=/home/pi/MagicMirror/mm.sh
Icon=/home/pi/Pictures/icon.png
Terminal=false
Type=Application
Categories=Utility;Application;
If you’ve left everything as the defaults when installing MM2 you wont need to change anything.
Save this file and place a copy of it in
/home/pi/.config/autostart
That’s it. Reboot and your mirror should load each time you reboot your Pi.
Definitely saves CPU usage thereby cutting down on power usage. There is also less to get “borked” up.
@n8ms Disables the onboard wireless. You might want to try bluetooth and a bluetooth keyboard. Why worry about the onboard? Just leave it.Doubt you are having power issues.
@n8ms add to config.txt
dtoverlay=pi3-disable-wifi
dtoverlay=pi3-disable-bt
(This also disables Bluetooth)
Adding a dongle on yours may be a good solution. use a usb extension cable yso you can move the dongle around if you dont get connectivity.
Facial Recognition
I wrote (with lots of help) a show video module which shows a video after a specific face is recognized. I would like to show a video if openCV recognizes any human standing in front of the camera. Also does anyone know the upper and lower values for the 3 algorithms? I’m having a very hard time recognizing individuals where the pictures used in training were taken with any camera other than the picam connected to the pi. I have no problems creating the XML file but when I run the test progran it doesn’t recognize the individuals whose face images were taken with the other cameras
@YourhighnessLeender Webcam is fine. If it worked in training it will work in the module.
@nuckinfutz behelper beat me to it.
@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…
Are you sure you put the training .xml file in the right place? Are you sure the training name syntax in your confis.jsg file matches the output from the train.py program you ran? those are about the only 2 things I can think of. Double check those and if you havent found the problem yet post the output from train.py and your config.js
There’s nothing wrong with the idea but why on earth would you title it “Discord”?
@sebzx10r What line did you put in the field? I can’t get it to work.