Identification of Most Suitable Binarisation Methods for Acehnese Ancient Manuscripts Restoration Software User Guide

Abstract

The Aceh Museum stores many digitized ancient manuscripts from hundreds of years ago. The condition of those manuscripts has degraded into several degradation types such as uneven contrast, show through effects, background spots, and text fading, which cause decreasing readability. A binarisation method is used to decrease the degradation effect on ancient manuscripts. Our research team is currently working on developing application software that consists of five binarisation methods, namely Otsu, Niblack, Sauvola, Lu, and Su for ancient manuscript restoration for the Aceh Museum staff to improve documents’ readability. In practice, a user still finds it difficult to choose the best method because there is no method that works best on every ancient manuscript for different types of degradation. This paper intends to determine a binarisation method that suits most manuscript conditions. The method used in this research includes the identification and classification of degradation types from 200 ancient Aceh digital manuscripts, followed by cropping the manuscripts to the size of 256 x 256 pixels. As many as five cropped areas from each degradation type are selected as research samples. These samples are binarisated using the methods. The last step is finding the most suitable binarisation method for each degradation type and classifying which methods are considered to have good readability, and that achieves at least 80% recall and precision values. From our experiments, we found that the Su binarisation methods demonstrate the best performance overall for every degradation type. Otsu, Lu, and Su are suited for uneven background; Sauvola, Lu, and Su are suited for showthrough effects; Otsu, Sauvola, and Su are suited for background spots; and Otsu and Su are suited for both text and background blurring and ‘fox’.

The Urgency of Digitizing Ancient Indonesian Manuscripts in Libraries: A Literature Review

“…Samsu & Sutriandi, 2016;Pageh et al, 2022;Hadi, Darsa, et al, 2017;Indiarti & Hasibin, 2019;Laras, 2021;Hasan, 2020;Laila & Wahdah, 2019; L. M. Samsu et al, 2020;Handoko, 2015;Kuswara, 2018;Jayanthi et al, 2017;Papaodysseus et al, 2014;Bhat & Seshikala, 2019;Kovalchuk, 2011;Hikmawati, 2016;Widyastuti & Saptono, n.d.;Maziyah, 2018;Abdel-Maksoud et al, 2023;Mohamed & Mohie, 2023;Ayu, 2019;Afriyanti, 2013;Rahayu, 2017;Suardiana, 2020;Suweta, 2019;Mulianawati et al, 2018;Buduroh, 2023;Fazarullah, 2016;Lolytasari & Istiqoriyah, 2018;Dewati et al, 2019;Taufiqurrahman & Hidayat, 2022. 3 The urgency of digitizing ancient manuscripts in Indonesia Handayani, 2023;Fakhriati et al, 2022;Cheriet et al, 2013;Čiuladienė et al, 2021;Savino et al, 2018;Doumat et al, 2010;Zakiyyah et al, 2022;Tajuddin et al, 2023;Rana et al, 2021;Siregar & Yahaya, 2023;Gaiani et al, 2022;Susilawati, 2017;...…”

“…Many cultural institutions in Indonesia possess ancient manuscripts that require preservation. For example, one museum with 1,400 manuscripts is undertaking such efforts (Susilawati, 2017), and the Aceh museum has started digitizing ancient manuscripts to prevent damage to the originals (Arnia et al, 2015). Digitalization reduces the handling of original manuscripts, which can lead to deterioration.…”

Moment invariant-based features for Jawi character recognition

“…Research in the digital image and computer vision field has yielded optical character recognition (OCR) for Chinese [5], Arabic [6], [7], Indian [8], and Latin characters [9], but no OCR system has yet been presented for Jawi characters. An OCR application framework requires four steps, namely: (1) preprocessing [10], [11], [12]; (2) segmentation; (3) feature extraction [13], [14]; and (4) pattern classiffiication [15]. At present, few algorithms for Jawi character recognition exist and these recognition algorithms have only been tested on a limited number of Jawi characters.…”

Improvement of binarization performance using local otsu thresholding

“…Several researchers tested the method for Jawi ancient document [13][14][15][16]. The result of this testing informed the existing method did not perform well for Jawi ancient document.…”