This repository contains:

 - Paper published in the Journal of data mining and digital humanities:
WasimEtAl_Toward_Automatic_Typography_Analysis__Serif_Classification_and_Font_Similarities.pdf

 - Two datasets: The first for serif classification 
 consisting of 126666 training and 2914 font-independent testing images in raster format. 
 The second dataset for svg based similarity learning consists of 124010 training and 2914 
 font-independent testing images. The images have been categorized into sans-serif, 
 linear-serif, slab-serif, and triangular serif by a designer at the EPFL+ECAL Lab

font_serif_dataset.zip
font_svg_dataset.zip

 - model weights:
serif_B_16_long_checkpoint.bin: refers to TransFG ViT B/16 Backend
serif_B_32_long_checkpoint.bin: refers to TransFG ViT B/32 Backend

ViT-B_16.npz ViT-B_32.npz are publicly available at: 
https://github.com/TACJu/TransFG
and were used for training TransFG ViT B/16 Backend and TransFG ViT B/32 Backend

 - to access the code used to perform the study please go to:
https://github.com/TalalWasim/GEST-Serif