PENGENALAN AKSARA INCUNG MENGGUNAKAN METODE HIDDEN MARKOV MODEL

Aksara Incung merupakan warisan budaya yang memerlukan upaya pelestarian melalui digitalisasi dan sistem pengenalan otomatis. Penelitian ini mengembangkan sistem pengenalan aksara Incung menggunakan metode Hidden Markov Model (HMM) dengan kombinasi ekstraksi fitur Intensity of Character (IoC) dan Mark Direction . Dataset terdiri dari 53 kelas aksara dengan 90 sampel citra per kelas. Evaluasi sistem dilakukan menggunakan k-fold cross validation (k=3 dan k=5) dengan variasi jumlah state 2 hingga 30. Hasil penelitian menunjukkan bahwa kombinasi HMM dengan ekstraksi ciri IoC 5x5 dan k-fold 5 menghasilkan akurasi terbaik sebesar 82.94%, sementara IoC 4x4 mencapai 82.49% dan IoC 3x3 mencapai 78.13%. Metode Mark Direction menghasilkan akurasi yang lebih rendah, dengan nilai 43.40% untuk arah vertikal dan 32.85% untuk arah horizontal. Penggunaan k-fold 5 secara konsisten memberikan hasil yang lebih baik dibandingkan k-fold 3, sementara jumlah state tidak menunjukkan pengaruh signifikan terhadap akurasi. Penelitian ini membuktikan efektivitas HMM dalam pengenalan aksara Incung, terutama ketika dikombinasikan dengan ekstraksi ciri IoC yang memiliki kompleksitas fitur lebih tinggi. Kata Kunci: Aksara Incung, Hidden Markov Model, Intensity of Character, Mark Direction, K-fold Cross Validation, Ekstraksi Ciri ABSTRACT The Incung script is a cultural heritage that requires preservation efforts through digitalization and automatic recognition systems. This research develops an Incung script recognition system using the Hidden Markov Model (HMM) method combined with Intensity of Character (IoC) and Mark Direction feature extraction. The dataset consists of 53 character classes with 90 image samples per class. System evaluation was conducted using k-fold cross validation (k=3 and k=5) with state variations ranging from 2 to 30. The results showed that the combination of HMM with 5x5 IoC feature extraction and k-fold 5 achieved the best accuracy of 82.94%, while 4x4 IoC achieved 82.49% and 3x3 IoC reached 78.13%. The Mark Direction method produced lower accuracy, with 43.40% for vertical direction and 32.85% for horizontal direction. The use of k-fold 5 consistently provided better results compared to k-fold 3, while the number of states showed no significant effect on accuracy. This research demonstrates the effectiveness of HMM in Incung script recognition, particularly when combined with IoC feature extraction that has higher feature complexity. Keywords: Incung Script, Hidden Markov Model, Intensity of Character, Mark Direction, K-fold Cross Validation, Feature Extraction.

Authors

Institutions

Publication Details

Journal
Syntax Journal of Software Engineering Computer Science and Information Technology
Published
2024-12-21
DOI
https://doi.org/10.46576/syntax.v5i2.5391
Primary Topic
Computer Science and Engineering
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

PENGENALAN AKSARA INCUNG MENGGUNAKAN METODE HIDDEN MARKOV MODEL

Kartono Pinaryanto, Andi Kristanto
Syntax Journal of Software Engineering Computer Science and Information Technology
Computer Science and Engineering
article

PENGENALAN AKSARA INCUNG MENGGUNAKAN METODE HIDDEN MARKOV MODEL

Kartono Pinaryanto, Andi Kristanto
article en

Abstract

Aksara Incung merupakan warisan budaya yang memerlukan upaya pelestarian melalui digitalisasi dan sistem pengenalan otomatis. Penelitian ini mengembangkan sistem pengenalan aksara Incung menggunakan metode Hidden Markov Model (HMM) dengan kombinasi ekstraksi fitur Intensity of Character (IoC) dan Mark Direction . Dataset terdiri dari 53 kelas aksara dengan 90 sampel citra per kelas. Evaluasi sistem dilakukan menggunakan k-fold cross validation (k=3 dan k=5) dengan variasi jumlah state 2 hingga 30. Hasil penelitian menunjukkan bahwa kombinasi HMM dengan ekstraksi ciri IoC 5x5 dan k-fold 5 menghasilkan akurasi terbaik sebesar 82.94%, sementara IoC 4x4 mencapai 82.49% dan IoC 3x3 mencapai 78.13%. Metode Mark Direction menghasilkan akurasi yang lebih rendah, dengan nilai 43.40% untuk arah vertikal dan 32.85% untuk arah horizontal. Penggunaan k-fold 5 secara konsisten memberikan hasil yang lebih baik dibandingkan k-fold 3, sementara jumlah state tidak menunjukkan pengaruh signifikan terhadap akurasi. Penelitian ini membuktikan efektivitas HMM dalam pengenalan aksara Incung, terutama ketika dikombinasikan dengan ekstraksi ciri IoC yang memiliki kompleksitas fitur lebih tinggi. Kata Kunci: Aksara Incung, Hidden Markov Model, Intensity of Character, Mark Direction, K-fold Cross Validation, Ekstraksi Ciri ABSTRACT The Incung script is a cultural heritage that requires preservation efforts through digitalization and automatic recognition systems. This research develops an Incung script recognition system using the Hidden Markov Model (HMM) method combined with Intensity of Character (IoC) and Mark Direction feature extraction. The dataset consists of 53 character classes with 90 image samples per class. System evaluation was conducted using k-fold cross validation (k=3 and k=5) with state variations ranging from 2 to 30. The results showed that the combination of HMM with 5x5 IoC feature extraction and k-fold 5 achieved the best accuracy of 82.94%, while 4x4 IoC achieved 82.49% and 3x3 IoC reached 78.13%. The Mark Direction method produced lower accuracy, with 43.40% for vertical direction and 32.85% for horizontal direction. The use of k-fold 5 consistently provided better results compared to k-fold 3, while the number of states showed no significant effect on accuracy. This research demonstrates the effectiveness of HMM in Incung script recognition, particularly when combined with IoC feature extraction that has higher feature complexity. Keywords: Incung Script, Hidden Markov Model, Intensity of Character, Mark Direction, K-fold Cross Validation, Feature Extraction.

Syntax Journal of Software Engineering Computer Science and Information TechnologyVol. 5(2)
Sanata Dharma University (ID)
Openalex Percentile: Top 77%
Computer Science and Engineering
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.