Research Article: Toward a Low-Resource Non-Latin-Complete Baseline: An Exploration of Khmer Optical Character Recognition

Research article on IEEE of the topic: "Toward a Low-Resource Non-Latin-Complete Baseline: An Exploration of Khmer Optical Character Recognition"

by RINA BUOY 1 , (Graduate Student Member, IEEE), MASAKAZU IWAMURA 1 , (Member, IEEE), SOVILA SRUN2 , AND KOICHI KISE1

ABSTRACT 

Many existing text recognition methods rely on the structure of Latin characters and words. Such methods may not be able to deal with non-Latin scripts that have highly complex features, such as character stacking, diacritics, ligatures, non-uniform character widths, and writing without explicit word boundaries. In addition, from a natural language processing (NLP) perspective, most non-Latin languages are considered low-resource due to the scarcity of large-scale data. This paper presents a convolutional Transformer-based text recognition method for low-resource non-Latin scripts, which uses local two dimensional (2D) feature maps. The proposed method can handle images of arbitrarily long text lines, which may occur with non-Latin writing without explicit word boundaries, without resizing them to a fixed size by using an improved image chunking and merging strategy. It has a low time complexity in self-attention layers and allows efficient training. The Khmer script is used as the representative of non-Latin scripts because it shares many features with other non-Latin scripts, which makes the construction of an optical character recognition (OCR) method for Khmer as hard as that for other non-Latin scripts. Thus, by analogy with the AI-complete concept, a Khmer OCR method can be considered as one of the non-Latin-complete methods and can be used as a low-resource non-Latin baseline method. The proposed 2D method was trained on synthetic datasets and outperformed the baseline models on both synthetic and real datasets. Fine-tuning experiments using Khmer handwritten palm leaf manuscripts and other non-Latin scripts demonstrated the feasibility of transfer learning from the Khmer OCR method. To contribute to the low-resource language community, the training and evaluation datasets will be made publicly available.

You can find full article at: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10316307 

Master Thesis of OCR using deep learning - Pavel Andrlik, 2022

Thanks for another citation by Pavel Andrlik, at the Master Thesis on OCR using deep learning.

University of West Bohemia
Faculty of Applied Sciences
Department of Cybernetics
(Czech Republic)

The author has been highlight my paper over the classification method choice on using Support Vector machine...[]

 Abstract of the Thesis

This diploma thesis deals with the problem of optical character recognition (OCR) using neural networks. I am focusing on improving text detection and OCR by fine-tuning an E2E-MLT scene text detector by training it on synthetic data which emulates real data. The model was fine-tuned on several datasets with synthetically generated data and real data, then the models were tested on one synthetic and two real datasets, one with the majority of the wild text, the second with the majority of TV news imprinted text. On the dataset with majority of TV news imprinted texts the fine-tuned models achieved improvement by decreasing character error rate from 52% to 31.6% word error rate and from 56.5% to 22%. It was also experimentally discovered that training models on synthetic data simulating real TV news images deteriorate detection and reading model capability on wild text data.

----------

What I am interesting is at the motivation side!

Full Thesis Can find at: https://otik.uk.zcu.cz/bitstream/11025/48953/1/Thesis___Pavel_Andrlik.pdf

My quick reflection on the motivation side!

The use case could also apply on some written paper for data collection such as on artist idea, random articles etc. we have a lot of handwriting or piece of writing printed that should also consider as collection on our language.