VerSign: Easy Signature Verification in Python
versign is a small Python package which can be used to perform verification of offline signatures.
It assumes no prior knowledge of any machine learning tools or machine learning itself, and therefore can be used by ML experts and anyone else who wants to quickly integrate this functionality into their project.
Getting Started
Installation
This package requires python 3. Installation can be done with pip:
pip install versign
You might also need to manually install the following dependencies:
pip install git+git://github.com/luizgh/visdom_logger#egg=visdom_logger
pip install git+https://github.com/luizgh/sigver.git
Installation inside a virtual environment is recommended.
Download Trained Models
Before you can get started with, there is one more step you need to complete. versign comes with some pre-trained models which give it its magic.
Download the compressed models here, and extract them to models/ directory in your project root. Your project directory should look something like this:
_ $PROJECT_ROOT
|__ models/
| |__ signet.pth
| |__ versign_segment.pkl
|__ ...
Organise Your Dataset
This model treats signature verification as a single-class learning problem where only positive samples (i.e. genuine signatures) are available during training. This is because, in real-world situations where we want to enrol users into a signature verification system for verifying their signatures later, we don't have any forgeries available unless we specifically obtain them. Which is not practical. However, both genuine and forged signatures can be present during testing.
Write Your First Program with VerSign
import os
from versign import VerSign
# Load training data
train_data # folder containing training data (only genuine samples)
x_train = [os.path.join(train_data, f) for f in sorted(os.listdir(train_data))]
# Load test data and labels
test_data # folder containing test data
x_test = [os.path.join(test_data, f) for f in sorted(os.listdir(test_data))]
# Train a writer-dependent model from training data
v = VerSign('models/signet.pth', (150, 220))
v.fit(x_train)
# Predict labels of test data
y_pred = v.predict(x_test)
For a more complete example and additional features such as measuring test accuracy if groundtruth is known, see example.py.
3 Aug 6, 2022
12 Jan 2, 2023
4 Mar 28, 2022
5 Oct 21, 2022
30 Sep 28, 2022
12 Dec 12, 2022
11 May 22, 2022
2 Dec 5, 2021
4.1k Dec 30, 2022
2 Feb 03, 2022
1 Oct 12, 2021
1 Jul 16, 2022
120 Dec 05, 2022
2 Nov 30, 2021
13 Dec 11, 2022
15 Aug 27, 2022
1 Jan 03, 2022
5 Jan 08, 2022
7 May 18, 2022
18 Nov 15, 2021
11 Aug 31, 2022
1 Nov 30, 2021
19 Jan 01, 2023
2 Dec 14, 2022
2 Feb 17, 2022
1 Dec 27, 2021
25 Aug 30, 2022
3 Nov 25, 2022