MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network

The lontar manuscript is an ancient Balinese cultural heritage written using Balinese characters on palm leaves. The recognition of Balinese characters in lontar is challenging because it has noise and limited data availability. To solve these problems, data augmentation is needed to increase the va...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Ni Putu Sutramiani, Nanik Suciati, Daniel Siahaan
Formato: article
Lenguaje:EN
Publicado: Elsevier 2021
Materias:
Acceso en línea:https://doaj.org/article/1eb1dcb8d5cc44b2b49090584e2410d1
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:1eb1dcb8d5cc44b2b49090584e2410d1
record_format dspace
spelling oai:doaj.org-article:1eb1dcb8d5cc44b2b49090584e2410d12021-11-30T04:16:42ZMAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network2405-959510.1016/j.icte.2021.04.005https://doaj.org/article/1eb1dcb8d5cc44b2b49090584e2410d12021-12-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S2405959521000497https://doaj.org/toc/2405-9595The lontar manuscript is an ancient Balinese cultural heritage written using Balinese characters on palm leaves. The recognition of Balinese characters in lontar is challenging because it has noise and limited data availability. To solve these problems, data augmentation is needed to increase the variety and amount of data to improve recognition performance. In this study, we collected Balinese character images from 50 lontar manuscript writers. We proposed MAT-AGCA that combines Adaptive Gaussian Thresholding and Convolutional Autoencoder for data augmentation. Based on experiments using InceptionResnetV2, DenseNet169, ResNet152V2, VGG19, and MobileNetV2, our proposed method achieved the best performance with 96.29% accuracy.Ni Putu SutramianiNanik SuciatiDaniel SiahaanElsevierarticleBalinese characterLontar manuscriptData augmentationAdaptive Gaussian ThresholdingConvolutional AutoencoderInformation technologyT58.5-58.64ENICT Express, Vol 7, Iss 4, Pp 521-529 (2021)
institution DOAJ
collection DOAJ
language EN
topic Balinese character
Lontar manuscript
Data augmentation
Adaptive Gaussian Thresholding
Convolutional Autoencoder
Information technology
T58.5-58.64
spellingShingle Balinese character
Lontar manuscript
Data augmentation
Adaptive Gaussian Thresholding
Convolutional Autoencoder
Information technology
T58.5-58.64
Ni Putu Sutramiani
Nanik Suciati
Daniel Siahaan
MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
description The lontar manuscript is an ancient Balinese cultural heritage written using Balinese characters on palm leaves. The recognition of Balinese characters in lontar is challenging because it has noise and limited data availability. To solve these problems, data augmentation is needed to increase the variety and amount of data to improve recognition performance. In this study, we collected Balinese character images from 50 lontar manuscript writers. We proposed MAT-AGCA that combines Adaptive Gaussian Thresholding and Convolutional Autoencoder for data augmentation. Based on experiments using InceptionResnetV2, DenseNet169, ResNet152V2, VGG19, and MobileNetV2, our proposed method achieved the best performance with 96.29% accuracy.
format article
author Ni Putu Sutramiani
Nanik Suciati
Daniel Siahaan
author_facet Ni Putu Sutramiani
Nanik Suciati
Daniel Siahaan
author_sort Ni Putu Sutramiani
title MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
title_short MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
title_full MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
title_fullStr MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
title_full_unstemmed MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network
title_sort mat-agca: multi augmentation technique on small dataset for balinese character recognition using convolutional neural network
publisher Elsevier
publishDate 2021
url https://doaj.org/article/1eb1dcb8d5cc44b2b49090584e2410d1
work_keys_str_mv AT niputusutramiani matagcamultiaugmentationtechniqueonsmalldatasetforbalinesecharacterrecognitionusingconvolutionalneuralnetwork
AT naniksuciati matagcamultiaugmentationtechniqueonsmalldatasetforbalinesecharacterrecognitionusingconvolutionalneuralnetwork
AT danielsiahaan matagcamultiaugmentationtechniqueonsmalldatasetforbalinesecharacterrecognitionusingconvolutionalneuralnetwork
_version_ 1718406790151929856