This repository contains the code and instructions to use the trained CNN models described in [1] to extract features for Offline Handwritten Signatures and use SVM models to train writer-dependent classifiers.
It also contains a web application which can predict the genuineness of an uploaded signature picture compared to the one used for SVM model training.
The web application frame is modified from https://github.com/sampathweb/apparel-styles/tree/master/app.
Create a new environment
conda create -n sigver -y python=3
source activate sigver
The following libraries are required
- Scipy version 0.18
- Pillow version 3.0.0
- OpenCV
- Pandas
- Theano
- Lasagne
- Tornado >= 4.4
- Scikit-learn >= 0.17
- requests >= 2.10
- jupyter >= 1.0
They can be installed by the following commands
conda install -y "scipy=0.18.0" "pillow=3.0.0"
pip install opencv-python
pip install pandas
pip install "Theano==0.9"
pip install https://github.com/Lasagne/Lasagne/archive/master.zip
pip install tornado
pip install scikit-learn
pip install requests
pip install jupyter
git clone https://github.com/EB324/signature_verification
cd signature_verification/sigver/models
wget "https://storage.googleapis.com/luizgh-datasets/models/signet_models.zip"
unzip signet_models.zip
- Put genuine signatures under
sigver/trainsig/genuine/ - Put forged signatures under
sigver/trainsig/forged/
The model is writer-dependent and can be trained for one writer each time only.
I have put some signatures (both genuine and forged) in them as a demo.
Create a temp-img folder
cd signature_verification/app/static/
mkdir temp-img
Run the App
python run_server.py
Open Browser: http://localhost:9000
Upload a signature and check its genuineness. You can use signatures in sigver/testsig/ as an example.
[1] Hafemann, Luiz G., Robert Sabourin, and Luiz S. Oliveira. "Learning Features for Offline Handwritten Signature Verification using Deep Convolutional Neural Networks" http://dx.doi.org/10.1016/j.patcog.2017.05.012 (preprint)
[2] Hafemann, Luiz G., Robert Sabourin, and Luiz S. Oliveira. "Fixed-sized representation learning from Offline Handwritten Signatures of different sizes" https://doi.org/10.1007/s10032-018-0301-6 (preprint)
[3] https://github.com/luizgh/sigver_wiwd
[4] https://github.com/sampathweb/apparel-styles/tree/master/app
GPDS: Vargas, J.F., M.A. Ferrer, C.M. Travieso, and J.B. Alonso. 2007. “Off-Line Handwritten Signature GPDS-960 Corpus.” In Document Analysis and Recognition, 9th International Conference on, 2:764–68. doi:10.1109/ICDAR.2007.4377018.