Gradient Pattern Analysis

Gradient Pattern Analysis (GPA) developed in Cython.

Requirements

Python 2.7 or greater
Numpy
Cython
Scipy

Install

pip install git+https://github.com/rsautter/GPA

Colab:

!pip install git+https://github.com/rsautter/GPA

Compilation

This code uses the Cython library, which transform pseudo-python code into C code. To get a better performance, edit compile.py and run:

python setup.py build_ext --inplace

The technique

Symmetry detection

Gradient Moments

Log

Oct. 7, 2022 - Updated Gradient Moments - new $G_1$ equation.
Apr. 18, 2021 - Added symmetrical and full analysis.
- now uses __call__ function to measure the gradient moments.
- vectors are now classified into symmetrical, asymmetrical and unknown.
Jun. 26, 2020 - Changed deploy system to setuptools

Jul. 01, 2019 - Changed GPA constructor, it receives modular and angular tolerance as input.
- Now the evaluation function receives the matrix as an argument.
- ScaleGPA has now a better performance.
- Now GPA works with double precision.
Nov. 27, 2018 - Added the FFT-Hilbert approach (file fragGPA.py)
Nov. 26, 2018 - Changed the user interface (it automatically identifies whether the file is a list or a data)
- Results are saved at reult.csv
- Tests updated

Mar. 27, 2018 - Added the Third and Fourth Gradient Moments ( $G_3$ and $G_4$ )

Aug. 30, 2017 - Added the First Gradient Moment ( $G_1$ )

References

https://en.wikipedia.org/wiki/Gradient_pattern_analysis

http://cython.org/