Studi Banding Implementasi Metode Hidden Markov Model dalam Pengenalan Tulisan Tangan

Muhammad Miftahul Amri

doi:10.56741/jgi.v1i01.26

Authors

Muhammad Miftahul Amri
muhammad.amri@te.uad.ac.id
Fakultas Teknologi Industri, Universitas Ahmad Dahlan https://orcid.org/0000-0001-5501-8484

DOI:

https://doi.org/10.56741/jgi.v1i01.26

Keywords:

Hidden Markov Model, Proses Tersembunyi, Pengenalan Tulisan, Proses Markov, Pengolahan Citra

Abstract

Hidden Markov Model (HMM) adalah model distribusi yang menghasilkan observasi yang bergantung pada keadaan pokok dan tidak dapat diamati (tersembunyi) pada Proses Markov. Metode pemodelan ini dapat diterapkan pada berbagai aplikasi secara fleksibel untuk data time series univariat dan multivariat, terutama untuk series dengan nilai diskrit, termasuk series pengelompokan dan series perhitungan. [1] HMM dapat digunakan untuk memprediksi kemungkinan demonstrasi massa, mengenali ekspresi wajah seseorang, seismokardiograf, dsb. Penggunaan metode HMM untuk pengenalan tulisan tangan telah sering dilakukan. Namun dari berbagai penelitian yang telah dilakukan, terdapat perbedaan metode ekstraksi fitur maupun metode pelatihan. Paper ini bertujuan untuk mengetahui metode dan parameter dari HMM yang menghasilkan performa terbaik dalam pengenalan tulisan tangan. Berdasarkan hasil perbandingan penelitian yang telah dilakukan sebelumnya [7, 8, 9, 10, 11, 12], didapatkan kombinasi metode yang paling optimal adalah ekstraksi 24 fitur meliputi kerapatan piksel hitam, posisi vertikal & horizontal piksel hitam, serta arah (kecekungan) piksel dengan batas ambang adaptif. Kemudian nilai tersebut diolah dengan menggunakan kombinasi planar HMM (hybrid) dan advanced HMM yang bersifat adaptif. Semakin banyak jumlah state dan sampel data, maka akurasi pengenalan akan semakin baik. Namun, waktu komputasi juga akan semakin lama. Maka, jumlah state dan sampel yang paling optimal perlu ditentukan melalui simulasi dan eksperimen.

Downloads

Download data is not yet available.

Author Biography

Muhammad Miftahul Amri, Fakultas Teknologi Industri, Universitas Ahmad Dahlan

Muhammad Miftahul Amri received the B.S. degree from the Department of Computer Science and Electronics, Universitas Gadjah Mada, Yogyakarta, Indonesia, in 2018, and the M.S. degree from the Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, South Korea, in 2021, where he is currently pursuing the Ph.D. degree. In 2021, he joined the faculty at Universitas Ahmad Dahlan, Indonesia, where he is currently a lecturer in the Department of Electrical Engineering. His research interests include wireless communications and artificial intelligence. He can be contacted at email: miftahulamri@ieee.org.

References

Zucchini, W, dan MacDonald, I. L, 2009, Hidden Markov Models for Time Series: An Introduction Using R, Chapman & Hall, Boca Raton.

Nasrollahi, S., Ebrahimi, A., 2013, Printed Persian sub-word recognition using wavelet packet descriptors, Hindawi Publishing Corporation Journal of Engineering, Vol 2013.

BayestehTashak, A., Ahmadifard, A.R., Khosravi, H., 2015, A two-step method for offline handwritten Farsi word recognition using adaptive division of gradient image, Signal and Data Processing, No.3, Vol.12.

Razavi, S.M., Kabir, E., 2004, Online Persian isolated character recognition, 6th Conference on Intelligent Systems, Kerman, Iran.

Razavi, S.M., Kabir, E., 2005, A simple recognition method for online isolated Farsi characters, Journal of School of Engineering, No.2, Vol.17, 65-82.

Natser, A.M., Aulama, M.M., Abandah, D. A. (2006)., Arabic Handwritten OCR Using HMM, Tesis, University of Jordan, Computer Engineering.

Elzobi, M., Al-Hamadi, A., Dings, L., Elmezain, M., Saeed, A., 2013, A Hidden Markov Model-based Approach with an Adaptive Threshold Model for Off-line Arabic Handwriting Recognition, 12th International Conference on Document Analysis and Recognition, hal 945-949, :http://www.ieeexplore.ieee.org/document/6628757/

Touj, S.M., Amara, N.E.B., Amiri, H., A Hybrid Approach for off-line Arabic Handwriting Recognition Based on a Planar Hidden Markov Modeling, 2008, 2008 International Conference on Signal and Image Processing, hal 684-689, :http://www.ieeexplore.ieee.org/document/4377058/

Xiang, D., Liu, H., Chen, X., Cheng, Y., Yao, H., 2012, Recognition of Off-line Arabic Handwriting using Hidden Markov Model Toolkit, 11th Internasional Symposium on Distributed Computing and Applications to Business, Enginering & Science, hal 409-412, :http://www.ieeexplore.ieee.org/document/7467702/

El-Hajj, R., Likforman-Sulem, L., Mokbel, C., 2005, Arabic Handwriting Recognition Using Baseline Dependant Features and Hidden Markov Modeling, 2005 Eight International Conference on Document Analysis and Recognition, hal 39-44, :http://www.ieeexplore.ieee.org/document/1575673

Hosny, I ., Abdou, S., Fahmy, A., 2011, Using Advanced Hidden Markov Models for Online Arabic handwriting recognition, 10th International Conference on Document Analysis and Recognition, hal 565-569, :http://www.ieeexplore.ieee.org/abstract/document/6166664

Feng, B., Ding, X., Wu, Y., Chinese Handwriting Recognition Using Hidden Markov Models, 2002 Five International Conference on Document Analysis and Recognition, hal 212-215, :http://www.ieeexplore.ieee.org/document/1047832/