Writer Retrieval and Writer Identification in Greek Papyri

TL;DR

This paper explores methods for writer identification and retrieval in Greek papyri, emphasizing the importance of binarization and feature sampling, and compares traditional, self-supervised, and supervised deep learning approaches.

Contribution

It introduces a focus on preprocessing and feature sampling for papyri, demonstrating that binarization significantly improves writer identification performance.

Findings

Binarization is crucial for effective writer identification in papyri.

Unsupervised feature methods perform comparably to supervised deep learning in writer re-identification.

Preprocessing steps significantly impact the accuracy of writer retrieval in challenging papyri data.

Abstract

The analysis of digitized historical manuscripts is typically addressed by paleographic experts. Writer identification refers to the classification of known writers while writer retrieval seeks to find the writer by means of image similarity in a dataset of images. While automatic writer identification/retrieval methods already provide promising results for many historical document types, papyri data is very challenging due to the fiber structures and severe artifacts. Thus, an important step for an improved writer identification is the preprocessing and feature sampling process. We investigate several methods and show that a good binarization is key to an improved writer identification in papyri writings. We focus mainly on writer retrieval using unsupervised feature methods based on traditional or self-supervised-based methods. It is, however, also comparable to the state of the art supervised deep learning-based method in the case of writer classification/re-identification.