Penggunaan Compressive Sensing pada Pengenalan Huruf dengan Tulisan Tangan Menggunakan Metode K-Nearest Neighbor

Penulis

  • Zalfa Alif Muhammad Universitas Telkom, Bandung
  • Gelar Budiman Universitas Telkom, Bandung
  • Sofia Saidah Universitas Telkom, Bandung

DOI:

https://doi.org/10.25126/jtiik.2022965612

Abstrak

Pengambilan, pengiriman, atau pertukaran informasi berupa data sangat dibutuhkan oleh manusia. Kehidupan manusia membutuhkan informasi dengan cepat dan tepat, maka data dimanfaatkan agar tidak memakan banyak kapasitas bandwidth dan memori saat sampai di server. Dalam penelitian ini, dilakukan kompresi dengan teknik CS (Compressive Sensing). Penggunaan CS difokuskan terhadap pengenalan huruf kapital dan angka tulisan tangan yang didapatkan dari sebelas mahasiswa dengan beberapa skenario. Untuk mengetahui hal tersebut dilakukan pengumpulan data lalu diolah menggunakan preprocessing, CS, rekontruksi dengan Orthogonal Matching Pursuit, dan proses terakhir pendeteksian menggunakan K-Nearest Neighbor yang didalamnya terdapat ekstraksi ciri menggunakan Template Matching. Setelah tahapan dirancang, dilakukan pengujian dengan beberapa skenario untuk memperoleh akurasi deteksi yang paling baik dengan mengubah parameter ukuran setiap blok pada suatu citra, mengubah baris kompresi, dan mengubah dimensi citra. Sehingga didapatkan bahwa skenario yang cocok untuk pengenalan huruf kapital dan angka adalah  skenario 1 (perbandingan database citra rekonstruktif pada pengujian dengan database citra asli sebelum akuisisi CS pada pelatihan) dengan akurasi deteksi sebesar 91.95% untuk huruf kapital sedangkan untuk deteksi angka sebesar 93%.

 

Abstract

Retrieval, delivery, or exchange of information in the form of data is needed by humans. Human life requires information quickly and precisely, so data are used so it doesn't take up a lot of bandwidth and memory capacity when it arrive at the server. In this study, compression was performed using the CS ( Compressive Sensing) technique. The use of CS is focused on recognizing capital letters and handwritten numbers obtained from eleven students with several scenarios. To find out, the data were collected and then processed using preprocessing, CS, reconstruction with Orthogonal Matching Pursuit, and the last detection process were using K- Nearest Neighbor in which there were feature extraction using Template Matching. After the design stage, several scenarios were tested to obtain the best detection accuracy by changing the size parameters of each block in an image, changing the compression line, and changing the image dimensions. So it is found that the suiTabel scenario for recognizing capital letters and numbers is scenario 1 ( comparison of the reconstructive image database in the test with the original image database before the acquisition of CS in training) with a detection accuracy of 91.95% for capital letters while for number detection it is 93%.


Downloads

Download data is not yet available.

Referensi

ANGRENI, I.A.A., ADISASMITA, S.A., RAMLI, M.I. dan HAMID, S., 2018. ‘Pengaruh Nilai K Pada Metode K-NEAREST NEIGHBOR (KNN) Terhadap Tingkat Akurasi Identifikasi Kerusakan Jalan’, Rekayasa Sipil, 7(2), pp. 63–70.

AULIANITA, R., 2016. ‘Komparasi Metode K-Nearest Neighbors dan Support Vector Machine Pada Sentiment Analysis Review Kamera’, Journal Speed – Sentra Penelitian Engineering dan Edukasi, 8(3), pp. 71–77.

FORNASIER, M. dan RAUHUT, H., 2015. ‘Compressive Sensing’, in Handbook of Mathematical Methods in Imaging: Volume 1, Second Edition, pp. 205–256.

MILLENIA, D., BUDIMAN, G. dan RIZAL, S., 2021. ‘Penerapan KNN Pada Deteksi ASCII Pasca Hasil Ekstraksi Audio Watermarking Menggunakan Teknik SMM Pada Segmen Audio Secara Adaptif’, e-Proceeding of Engineering Telkom University, 8(2), pp. 1497–1504. Tersedia di: <https://openlibrarypublications.telkomuniversity.ac.id/index.php/engineering/article/view/14557> [Diakses 10 Mei 2021]

MOSTAFA, M., & FAKHR, M.W., 2017. ‘Joint Image Compression and Encryption Based on Compressed Sensing and Entropy Coding’, Proceedings - 2017 IEEE 13th International Colloquium on Signal Processing and its Applications, CSPA 2017, pp. 129–134.

NUGROHO, C. A., 2021. ‘Klasifikasi K-Nearest Neighbor Chest X-Ray Pasien Covid-19 dengan Haralick Features dan Histogram of Oriented Gradient’, Jurnal Ilmiah Matematika, 9(1), pp. 188–195. Tersedia di: <https://www.neliti.com/id/publications/270379/pengaruh-nilai-k-pada-metode-k-nearest-neighbor-knn-terhadap-tingkat-akurasi-ide> [Diakses 7 Januari 2021]

PRAMUDANA, Y., 2015. ‘Pengenalan Tulisan Tangan Dengan Menggunakan Metode Diagonal Feature Extraction dan K-Nearest Neighbour’, eProceedings of Engineering Telkom University, 2, pp. 1–12. Tersedia di: <https://openlibrarypublications.telkomuniversity.ac.id/index.php/engineering/article/view/1192> [Diakses 27 November]

SALAHDINE, F., KAABOUCH, N. dan GHAZI, H. El., 2018. ‘One-Bit Compressive Sensing Vs. Multi-Bit Compressive Sensing for Cognitive Radio Networks’, Proceedings of the IEEE International Conference on Industrial Technology, (1), pp. 1610–1615.

SUN, H. dan NI, L., 2013. ‘Compressed Sensing Data Reconstruction Using Adaptive Generalized Orthogonal Matching Pursuit Algorithm’, pp. 1102–1106.

SUPARDI, J., HAPSARI, I. A. dan SIRAJ, M. M., 2014. ‘Handwritten Alphabets Recognition Using Twelve Directional Feature Extraction and Self Organizing Maps’, Proceeding - 2014 International Conference on Computer, Control, Informatics and Its Applications: ‘New Challenges and Opportunities in Big Data’, IC3INA 2014, pp. 149–153.

TRIPATHY, S. dan PANDA, G., 2016. ‘A Novel Classification of Handwritten Digits Using Compressive Sensing Technique’, in 2016 International Conference on Computational Techniques in Information and Communication Technologies, ICCTICT 2016 - Proceedings, pp. 233–236.

USMAN, K., 2017. Introduction to Orthogonal Matching Pursuit, Telkom University. Tersedia di: <https://korediantousman.staff.telkomuniversity.ac> [Diakses 10 November 2020]

WANG, J.C., LEE, Y.S., LIN, C.H., WANG, S.F., SHIN, C.H. dan WU, C.H., 2016. ‘Compressive Sensing-Based Speech Enhancement’, IEEE/ACM Transactions on Audio Speech and Language Processing, 24(11), pp. 2122–2131.

WIRASNO, M. H., OSMOND, A.B. dan PRASASTI, A. L., 2019. ‘Pengenalan Pembuluh Darah Jari Manusia Untuk Autentikasi Dengan Maximum Curvature Points Segmentation Dan Template Matching’, e-Proceeding of Engineering Telkom University, 6(3), pp. 10332–10340. Tersedia di: <https://openlibrarypublications.telkomuniversity.ac.id/index.php/engineering/article/view/11185> [Diakses 6 April 2021]

ZOBLY, S. M. S. dan KADAH, Y. M., 2012. ‘Orthogonal Matching Pursuit & Compressive Sampling Matching Pursuit for Doppler Ultrasound Signal Reconstruction’, 2012 Cairo International Biomedical Engineering Conference (CIBEC), pp. 52–55.

Diterbitkan

22-12-2022

Terbitan

Bagian

Ilmu Komputer

Cara Mengutip

Penggunaan Compressive Sensing pada Pengenalan Huruf dengan Tulisan Tangan Menggunakan Metode K-Nearest Neighbor. (2022). Jurnal Teknologi Informasi Dan Ilmu Komputer, 9(6), 1237-1246. https://doi.org/10.25126/jtiik.2022965612