Modified Segmentation Algorithm for Recognition of Older Geez Scripts Written on Vellum

TL;DR

This paper presents a modified segmentation algorithm combined with adaptive filtering, thresholding, and bounding box projection to improve recognition of ancient Geez scripts on vellum, achieving nearly 80% accuracy.

Contribution

The study introduces a novel segmentation approach specifically designed for recognizing older Geez scripts on vellum, enhancing accuracy over previous methods.

Findings

Achieved 79.32% recognition accuracy with the proposed method.

Effective noise reduction using adaptive filtering.

Improved segmentation of strokes between characters, numbers, and punctuation.

Abstract

Recognition of handwritten document aims at transforming document images into a machine understandable format. Handwritten document recognition is the most challenging area in the field of pattern recognition. It becomes more complex when a document was written on vellum before hundreds of years, like older Geez scripts. In this study, we introduced a modified segmentation approach to recognize older Geez scripts. We used adaptive filtering for noise reduction, Isodata iterative global thresholding for document image binarization, modified bounding box projection to segment distinct strokes between Geez characters, numbers, and punctuation marks. SVM multiclass classifier scored 79.32% recognition accuracy with the modified segmentation algorithm.