Benchmarking of Document Image Analysis Tasks for Palm Leaf Manuscripts from Southeast Asia

Abstract

This paper presents a comprehensive test of the principal tasks in document image analysis (DIA), starting with binarization, text line segmentation, and isolated character/glyph recognition, and continuing on to word recognition and transliteration for a new and challenging collection of palm leaf manuscripts from Southeast Asia. This research presents and is performed on a complete dataset collection of Southeast Asian palm leaf manuscripts. It contains three different scripts: Khmer script from Cambodia, and Balinese script and Sundanese script from Indonesia. The binarization task is evaluated on many methods up to the latest in some binarization competitions. The seam carving method is evaluated for the text line segmentation task, compared to a recently new text line segmentation method for palm leaf manuscripts. For the isolated character/glyph recognition task, the evaluation is reported from the handcrafted feature extraction method, the neural network with unsupervised learning feature, and the Convolutional Neural Network (CNN) based method. Finally, the Recurrent Neural Network-Long Short-Term Memory (RNN-LSTM) based method is used to analyze the word recognition and transliteration task for the palm leaf manuscripts. The results from all experiments provide the latest findings and a quantitative benchmark for palm leaf manuscripts analysis for researchers in the DIA community.

AKSALont: Automatic transliteration application for Balinese palm leaf manuscripts with LSTM Model

“…Sistem tidak akan mampu melakukan rekognisi teks pada teks kata baru yang sebelumnya tidak ada pada dataset klasifikasi. Upaya pembangunan mesin transliterasi teks Lontar Bali juga telah dilakukan dalam [10] yang menggunakan framework pengenal teks OCRopy [11]. Pada penelitian tesebut, uji performansi transliterasi tidak hanya dilakukan dengan dataset Lontar Bali, tetapi juga dengan dataset Lontar Sunda [12] dan Kamboja [3].…”

“…Tantangan kompleksitas aksara dan keragaman kondisi kerusakan citra Lontar Bali sebagai bagian dari tantangan teknis di penelitian ini memang cukup tinggi Error: Reference source not found. Di samping itu, kuantitas data latih dan proporsi jumlah masing-masing sampel bentuk aksara Bali beserta kombinasi bentuk aksara dalam penulisannya yang terdapat dalam data latih juga turut mempengaruhi performansi hasil pembelajaran model LSTM [9], [10]. Hasil yang dicapai sampai saat ini merupakan hasil tahap uji yang sudah cukup baik, dan sudah mampu menunjukkan bahwa sistem bisa mentransliterasi citra digital aksara Bali.…”

A Scheme Towards Automatic Word Indexation System for Balinese Palm Leaf Manuscripts

“…The main contribution of this work lies in the fact that this system is supported through our in-depth study results on various components of the document image analysis (DIA) system for lontar collections [1][2][3][4][5][6][7].…”

Image Patch Extraction in Text Area of Balinese Palm Leaf Manuscripts with Gabor Filters

“…To deal with those problems, a preliminary project to digitize the collections of Lontar and to build an automatic transliteration machine to convert the Balinese script of Lontar into Roman script has been done [1], [2]. These previous works were proposed to support and to complete the final goal of building an automatic indexing system for Lontar collections.…”