Thank you Dennis, you provided us a collection of 40 males each 10 pics (100×100 pixels).
He wrote in its comments “Now I get much much better results. I thought I share the pics and csv file with you (csv with that many pics is quite some work).Oh, the csv file has only 50 pics, the facerec software could not handle 400 of them. I works nicely with 50. You have to edit the cpp file, I changed lise in “others” and display a text “unidentified other male person”. Enjoy!”
Zip file can be download here : http://raufast.org/download/100×100.zip
Thank you !
It’s finaly working 🙂 The whole thing took me about 8 hours to work mostly because of so many minor corrections.
Could you please upload the Images to another host?
This one is crap since I had to “register” with a social platform account.
@philipp (and others) here is the file hosted on dropbox: (no reg required)
Paulo Pine said:
Pierre, do you have any good idea how to get the video stream over web? Like we do using “motion”? Great Job.
I downloaded the 100×100.zip and put it in my camcv directory. I changed ‘csv.txt’ to faces.txt’
Here’s the problem when I run ./camcv, I am getting this output:
./camcv faces.csv 1 5000
(init) People initialized
(init) 400 pictures read to train
(init) 0 pictures of others (0) read to train
(init) taille images ok
(init) start train images
It looks like it is getting stuck at start train images. How can I fix this?
it takes ALOT of time to train mine took about 4-5 minites and i have overclocked RPI a little…. patience i think is solution in your case
Paulo Pine said:
” else if (!raspipreview_create(&state.preview_parameters)) ” line in main to
” else if ( (status = raspipreview_create(&state.preview_parameters)) != MMAL_SUCCESS) ”
Can somebody send back the zip file ?
Thanks in advance
Yes, I’ll do it tonight
done. Follow the new link
Thank you Pierre.
I used the 400 pictures and got this error:
OpenCV error: Image step is wrong (The matrix is not continuous, thus its number of rows can not be changed). Thanks.
Setiawan Qodri Nugroho said:
Can somebody send back the zip file ? Link Not Found.
Thanks in advance
Same problem here with zip file 🙂
Samuel Pickard said:
The original link at the top of the article still works.
This is probably a shot in the dark, but I need some help with this portion of the project. So I got everything working fine, openCV, face detection, but the recognition algorithm has some glitches. I captured about 30 photos of myself and created the face.csv file that points to my training photos. I don’t have any photos for anyone else (for the time being I would just like to get the recognition program to work for my face) I do have the collection of photos from the zip file above, but im not sure of how to setup the program to use them as negative photos. So my question is: how do I setup these negative images in the face.csv file so that the program can better face recognize faces. Any help will be appreciated. Thanks
@sq7: If you still need help with the tons of negative photos for training. All you have to do is to copy them all (if u want to use all) into the directory/folder where your photos are located and give each a unique name (say 0xx.jpg). Then using the same convention you are currently employing for your photos, define the a link for each photo in the faces.txt file (e.g. /home/pi/faces/0xx.jpg). The rest is easy…