Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learning based Ge'ez character handwritten recognition

Hailemicael Lulseged Yimer, Hailegabriel Dereje Degefa, Marco Cristani, Federico Cunico

TL;DR

A dual-stage recognition approach achieves new top scores in Ge’ ez handwriting recognition, outperforming eight state-of-the-art methods, which are SVTR, ASTER, and others as well as human performance, as measured in the HHD- Ethiopic dataset work.

Abstract

Ge'ez, an ancient Ethiopic script of cultural and historical significance, has been largely neglected in handwriting recognition research, hindering the digitization of valuable manuscripts. Our study addresses this gap by developing a state-of-the-art Ge'ez handwriting recognition system using Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. Our approach uses a two-stage recognition process. First, a CNN is trained to recognize individual characters, which then acts as a feature extractor for an LSTM-based system for word recognition. Our dual-stage recognition approach achieves new top scores in Ge'ez handwriting recognition, outperforming eight state-of-the-art methods, which are SVTR, ASTER, and others as well as human performance, as measured in the HHD-Ethiopic dataset work. This research significantly advances the preservation and accessibility of Ge'ez cultural heritage, with implications for historical document digitization, educational tools, and cultural preservation. The code will be released upon acceptance.

Learning based Ge'ez character handwritten recognition

TL;DR

A dual-stage recognition approach achieves new top scores in Ge’ ez handwriting recognition, outperforming eight state-of-the-art methods, which are SVTR, ASTER, and others as well as human performance, as measured in the HHD- Ethiopic dataset work.

Abstract

Ge'ez, an ancient Ethiopic script of cultural and historical significance, has been largely neglected in handwriting recognition research, hindering the digitization of valuable manuscripts. Our study addresses this gap by developing a state-of-the-art Ge'ez handwriting recognition system using Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. Our approach uses a two-stage recognition process. First, a CNN is trained to recognize individual characters, which then acts as a feature extractor for an LSTM-based system for word recognition. Our dual-stage recognition approach achieves new top scores in Ge'ez handwriting recognition, outperforming eight state-of-the-art methods, which are SVTR, ASTER, and others as well as human performance, as measured in the HHD-Ethiopic dataset work. This research significantly advances the preservation and accessibility of Ge'ez cultural heritage, with implications for historical document digitization, educational tools, and cultural preservation. The code will be released upon acceptance.

Paper Structure

This paper contains 9 sections, 2 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related work
  3. Datasets
  4. Handwritten character recognition
  5. Method
  6. Architecture
  7. Implementation details
  8. Experiments and Results
  9. Conclusion

Figures (4)

  • Figure 1: Example of a Ge'ez handwritten manuscript. This historical document exemplifies the challenges in optical character recognition (OCR) for low-resource scripts, highlighting the intricate and variable nature of Ge'ez handwriting. It underscores the necessity of advanced machine learning and deep learning techniques to accurately digitize and preserve such culturally significant texts
  • Figure 2: Challenges in Ge'ez handwriting recognition. (a) and (b) demonstrate the variation in handwriting styles for the same word and characters. As visible, the difference in appearance is significant, highlighting the complexity of recognition across different scribes or historical periods. (c) shows a subset of the 182 Ge'ez characters, focusing on those that are particularly difficult to distinguish due to their similar forms.
  • Figure 3: Examples of predictions made by the model. (a) Character-level predictions. (b) Text-level predictions.
  • Figure 4: The loss and classification accuracy of the character recognition model during training. The system seems to learn the task in a few epochs, and further training did not significantly improve accuracy.