Segementasi Gambar pada Dataset MNIST dengan Optimasi Mini Batch dan K-means++ pada Algoritma K-means

Agus Susanto, Rusdianto Roestam

Abstract


Proses segmentasi gambar tulisan tangan berupa gambar dengan menggunakan k-means perlu diawali dengan clusterisasi. Akibat dari proses clusterisasi ini membuat proses secara keseluruhan menjadi lambat. Penelitian ini memanfaatkan mini batch k-means untuk mempercepat proses pengenalan tulisan tangan berupa gambar. Data yang digunakan pada penelitian ini adalah dataset MNIST. Normalisasi terhadap dataset akan dilakukan terlebih dahulu sebelum clusterisasi menggunakan teknik minibatch k-means diterapkan. Selanjutnya inisiasi centroid dilakukan menggunakan k-means++. Hasil penelitian menyatakan bahwa optimasi minibatch dan k-means++ pada algoritma k-means mampu mempercepat waktu komputasi lebih singkat 28 menit 13 detik yaitu 29 menit 39 detik menjadi 1 menit 26 detik, dengan akurasi sebesar 90,13%.


Keywords


Segmentasi Gambar; Clustering; k-means; mini batch; k-means++; MNIST

Full Text:

PDF

References


Z. Ge, Z. Song, S. X. Ding, and B. Huang, “Data Mining and Analytics in the Process Industry: The Role of Machine Learning,” IEEE Access, vol. 5, pp. 20590–20616, 2017, doi: 10.1109/ACCESS.2017.2756872.

J. Gan, W. Wang, and K. Lu, “Compressing the CNN architecture for in-air handwritten Chinese character recognition,” Pattern Recognit. Lett., vol. 129, pp. 190–197, 2020, doi: 10.1016/j.patrec.2019.11.028.

A. F. Agarap and A. P. Azcarraga, “Improving k-Means Clustering Performance with Disentangled Internal Representations,” Proc. Int. Jt. Conf. Neural Networks, no. 2016, 2020, doi: 10.1109/IJCNN48605.2020.9207192.

W. Jiang, “MNIST-MIX: a multi-language handwritten digit recognition dataset,” IOP SciNotes, vol. 1, no. 2, p. 025002, 2020, doi: 10.1088/2633-1357/abad0e.

M. Vafadar et al., “Unsupervised K-means clustering algorithm,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 8, no. 3, pp. 1–12, 2019, doi: 10.1109/ACCESS.2020.2988796.

N. Dhanachandra, K. Manglem, and Y. J. Chanu, “Image Segmentation Using K-means Clustering Algorithm and Subtractive Clustering Algorithm,” Procedia Comput. Sci., vol. 54, pp. 764–771, 2015, doi: 10.1016/j.procs.2015.06.090.

Soeren Bergmann Niclas Feldkamp and S. Strassburger, “APPROXIMATION OF DISPATCHING RULES FOR MANUFACTURING SIMULATION USING DATA MINING METHODS,” Dk, vol. 53, no. 9, pp. 1689–1699, 2015, doi: 10.1017/CBO9781107415324.004.

V. Kathiresan and P. Sumathi, “An efficient clustering algorithm based on Z-Score ranking method,” 2012 Int. Conf. Comput. Commun. Informatics, ICCCI 2012, pp. 10–13, 2012, doi: 10.1109/ICCCI.2012.6158779.

K. Peng, V. C. M. Leung, and Q. Huang, “Clustering Approach Based on Mini Batch Kmeans for Intrusion Detection System over Big Data,” IEEE Access, vol. 6, pp. 11897–11906, 2018, doi: 10.1109/ACCESS.2018.2810267.

G. S. Thejas, S. R. Joshi, S. S. Iyengar, N. R. Sunitha, and P. Badrinath, “Mini-Batch Normalized Mutual Information: A Hybrid Feature Selection Method,” IEEE Access, vol. 7, no. Mi, pp. 116875–116885, 2019, doi: 10.1109/ACCESS.2019.2936346.

S. R. Fitriyani and H. Murfi, “The K-means with mini batch algorithm for topics detection on online news,” 2016 4th Int. Conf. Inf. Commun. Technol. ICoICT 2016, vol. 4, no. c, 2016, doi: 10.1109/ICoICT.2016.7571914.

D. Arthur and S. Vassilvitskii, “K-means++: The advantages of careful seeding,” Proc. Annu. ACM-SIAM Symp. Discret. Algorithms, vol. 07-09-Janu, pp. 1027–1035, 2007.

S. Ahlawat and A. Choudhary, “Hybrid CNN-SVM Classifier for Handwritten Digit Recognition,” Procedia Comput. Sci., vol. 167, no. 2019, pp. 2554–2560, 2020, doi: 10.1016/j.procs.2020.03.309.

Y. Wang, “Research on Moment of Inertia Measurement Method,” IOP Conf. Ser. Mater. Sci. Eng., vol. 592, no. 1, 2019, doi: 10.1088/1757-899X/592/1/012078.

O. Faker and E. Dogdu, “Intrusion detection using big data and deep learning techniques,” ACMSE 2019 - Proc. 2019 ACM Southeast Conf., pp. 86–93, 2019, doi: 10.1145/3299815.3314439.




DOI: http://dx.doi.org/10.26418/justin.v11i2.55503

Refbacks

  • There are currently no refbacks.


Copyright (c) 2023 JUSTIN (Jurnal Sistem dan Teknologi Informasi)

ara komputer
View My Stats

Creative Commons License
All article in Justin is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License