panther: A Python repository from Algorithmic and Statistical Information Laboratory - Algorithmic and Statistical Information Laboratory

Measuring probabilistic-algorithmic information of artistic paintings

INSTALL

Get PANThER project using:

git clone https://github.com/asilab/panther.git
cd panther/

RUN

Give run permissions to the files:

chmod +x *.sh
bash make.sh
pip3 install -r requirements.txt

Furthermore, use the following instructions.

To run the pipeline and obtain all the Reports in the folder reports, use the following command in the src dir:

./Run.sh

This will run the following scripts automatically:

./Dataset.sh                        # Downloads and unzips dataset
./Quantize.sh                       # Quantizes images of the dataset to 8, 6, 4 and 2 bits.
./normalize_images.sh               # Normalizes 0 to 256 the 8 bit images.
./Trimm_and_Binarization.sh         # Trims and Binarizes images of the dataset.
./BDM.sh                            # Computes NBDM (1 and 2) for all quantized images of the dataset.
./Compress.sh                       # Computes NC for all quantized images of the dataset.
./HDC.sh                            # Computes HDC alpha for 8-bit quantized images of the dataset.
./Average_Complexity.sh             # Computes average information-based measures for each author
./Region_Complexity.sh              # Computes regional NC for 8-bit quantized images of the dataset.
./Average_Regional_Complexity.sh    # Computes fingerprint of each author

To download and prepare the dataset, use the following command:

./Dataset.sh

To benchmark the compressors, use the following command:

./Benchmark.sh

To quantitize images run, to trim and binarize, use the following command:

./Quantize.sh                     
./Trimm_and_Binarization.sh

To perform comparisson between NC, NBDM1 and NBDM2, use the following command:

./Compare.sh

To compute the average NC, NBDM1, and NBDM2 for each author, use the following command:

./Average_Complexity.sh

To compute the NC with the HDC results, use the following command:

./NC_HDC.sh

To recreate the reports of Regional Complexity, use the following command:

./Region_Complexity.sh

To recreate the reports of fingerprints, use the following command:

./Average_Regional_Complexity.sh

To recreate the authors' fingerprints, use the following command:

./Fingerprints.sh

To recreate the phylogenic trees, use the following command:

./Tree.sh

To assess the author average variation and percentage difference between normalized and non-normalized measures, use the following command:

./norm_vs_non_norm.sh

To perform the Mantel test and view the average variance between different distance matrices, use the following command:

./Mantel_test_and_variation.sh

To perform author classification, run the jupyter file:

Painting91_author_classification.ipynb

To perform style classification, run the jupyter file

Painting91_style_classification.ipynb

To assess the normality properties, use the following command:

./Idempotency.sh
./Symmetry.sh
./Triangular.sh

CITE

Please cite the followings, if you use PANThER:

Silva, Jorge Miguel, et al. "Automatic analysis of artistic paintings using information-based measures." Pattern Recognition 114 (2021): 107864.

BibTex

@article{silva2021automatic,
  title={Automatic analysis of artistic paintings using information-based measures},
  author={Silva, Jorge Miguel and Pratas, Diogo and Antunes, Rui and Matos, S{\'e}rgio and Pinho, Armando J},
  journal={Pattern Recognition},
  volume={114},
  pages={107864},
  year={2021},
  publisher={Elsevier}
}

RELEASES

Release 1.

ISSUES

Please let us know if there is any issues.

LICENSE

PANThER is under GPL v3 license. For more information, click here.