ee09115 / spoofing_detection Goto Github PK

Hi,
Thank you for sharing your great work!

Do you have the datasets that you experiemented with in your works? Particularly OULU-NPU and Idiap Replay-Attack datasets, since the owners of these datasets may no longer maintain their website. It would be nice if you can share them!

Thanks

if np.mean(measures) >= 0.7:

Why use the value 0.7?

Hi, thanks for your excellent model.
I'm trying to make an anti-spoofing model myself and I was wondering if you could share the preprocess you have made to the images

Hey guys. first of all, well done on the project. And thanks for publishing it for others to play with.

I'm fairly new to python and I was hoping that you might be able to tell me what params I need to give the (Main.py) script when I run it.

Or maby give like an example of how I can run this using windows CMD.

Thanks in advance.
Kind Regards
Ben-Ron Geldenhuys

What parameters need to be entered in: --name, --device, --threshold? And how can I debug?

Thanks!!

How to setup this on ubantu or windows 10?

training.py needs csv file as argument .Where is it ? actually what is it ? İs it dataset ?and which columns?

Hi,

i would like to produce some own datasets with webcam and own print magazins.

Do you can show us how you train the models by providing the training script?

Thx.

what is the best performance value for threshold argument ?

Edit: Finded graphics at your folders sorry mate.

I found out that your model only works perfectly under the conditions where the face's brightness is normal or dark. I tried some real faces(like the attached photo) under light (a very normal case) and it was not successful. Do you have any solution for that problem?

help me please!

it's wrong
https://i.pinimg.com/originals/b2/66/55/b2665549eb15a7656367f9423935d2dd.png

Hello again,

Thanks for sharing your work.

I read a few issues and im thinking this may not work well on "real world" data, so i wanted to know how could i make my own dataset, given n vidoes which are real and m videos which are attacked in some way.It is related to my other question of how to compute the feature vector given an image.

Thanks in advance.

no one can legal use your awesome code in their projects now. please add MIT license file

Unable to open the models using joblib or pickle..
It always looks for sklearn.externals.joblib

How to get the models saved in using joblib instead of scikit-externals.joblib

Can anyone please share detailed instructions to use this code. I am not able to run this code. How to detect spoofing from webcam or not video using this code?

Thank you in advance.

Hi thanks for providing your hardwork.

But results needs to be improved, do you have any suggestions for further performance improvement.

thanks

is there any tutorial on how can i run this project ??

Can you provide the CSV file training data used for training_script.py?
joblib could not load the classifiers in Raspberry Pi 3B+ due to scikit-learn incompatible versions.
I need to use scikit-learn 0.20.2 instead of 0.19.1 so I need to train the classifier with this version.

Thank you!

I just run the program with camera.

hi:
your job is good,but i find that real face do not be detected . can u give some good pre model to test your job

hi,thank for you share your code. how you plot this . Can I get a drawing script?

Hi valter costa, I checked out your project but I am not able to download any database and having some error while running it, can you provide me a zip file that contains all the files of the project including the databases? that would be really helpful.Thanks