Pristanti, Yuliana Diah (2018) Identifikasi Tanda Tangan dengan Ekstraksi Ciri GLCM dan LBP. Magister thesis, Universitas Brawijaya.
Abstract
Identifikasi tanda tangan menggunakan ekstraksi fitur GLCM (The Grey Level Co-occurrence Matrix) dan LBP (The Local Binary Pattern) dengan membandingkan hasil akurasi keduanya. Data tanda tangan yang digunakan dengan tinta hitam dari 15 orang, dimana masing-masing terdapat 10 tanda tangan. Data tanda tangan bertinta warna selain hitam dari 5 orang, masing-masing dengan 10 tanda tangan. Untuk data pelatihan diambil 7 tanda tangan dari masing-masing orang sehingga data latih berjumlah 105 tanda tangan untuk bertinta hitam dan 45 tanda tangan untuk bertinta warna selain hitam. Sedangkan data pengujian diambil 3 tanda tangan dari masing-masing orang sehingga data uji berjumlah 45 tanda tangan bertinta hitam dan 15 tanda tangan bertinta warna selain hitam. Dari hasil pengolahan citra didapatkan prosentase menggunakan ekstraksi ciri GLCM lebih besar dibandingkan prosentase menggunakan ekstraksi ciri LBP, yaitu GLCM mencapai 86,67% dan LBP 80,00% dengan tanda tangan bertinta hitam. Sedangkan hasil akurasi jika ditambahkan data tanda tangan bertinta warna selain hitam, untuk GLCM sebesar 80,00% dan LBP 78,33%. Namun keduanya tetap berada pada tingkat keberhasilan yang tinggi. Sehingga bisa disimpulkan bahwa kedua ekstraksi ciri, baik GLCM maupun LBP bisa direkomendasikan untuk mengenali tekstur tanda tangan bertinta hitam maupun warna.
English Abstract
Identify signatures using extraction features of The Gray Level Co-occurrence Matrix (GLCM) and LBP (The Local Binary Pattern) by comparing the results of both accuracy. Signature data used in black ink from 15 people, each of which has 10 signatures. Data other than black color inked signatures of 5 people, each with 10 signatures. For the training data, 7 signatures were taken from each person so that the training data was 105 signatures for black ink and 45 signatures for color ink other than black. While the test data was taken 3 signatures from each person so that the test data amounted to 45 black inked signatures and 15 color inked signatures other than black. From the results of image processing, the percentage using GLCM feature extraction was greater than the percentage using LBP feature extraction, is GLCM reached 86.67% and LBP 80.00% with black inked signatures. Whereas the results of the accuracy if added with data on color inked signatures other than black, for GLCM is 80.00% and LBP 78.33%. But both remain at a high success rate. So it can be concluded that both feature extraction, both GLCM and LBP, can be recommended for recognizing black and color inked signature textures.
Other obstract
-
| Item Type: | Thesis (Magister) |
|---|---|
| Identification Number: | TES/006.24/PRI/i/2018/041811857 |
| Uncontrolled Keywords: | AUTOMATIC IDENTIFICATION SYSTEMS |
| Subjects: | 000 Computer science, information and general works > 006 Special computer methods > 006.2 Special-purpose systems > 006.24 Automatic identification and data capture (AIDC) |
| Divisions: | S2/S3 > Magister Teknik Elektro, Fakultas Teknik |
| Depositing User: | Budi Wahyono Wahyono |
| Date Deposited: | 05 Sep 2019 06:41 |
| Last Modified: | 22 Apr 2020 16:39 |
| URI: | http://repository.ub.ac.id/id/eprint/172429 |
Actions (login required)
 |
View Item |