Syafiq, Muhammad (2016) Rancang bangun sistem pengenal biometris pembuluh darah vena pada telapak tangan menggunakan cahaya inframerah. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2412100057-undergraduate-thesespdf.pdf Download (200kB) | Preview |
Preview |
Text
2412100057-paperpdf.pdf Download (178kB) | Preview |
Preview |
Text
2412100057-presentationpdf.pdf Download (1MB) | Preview |
Abstract
Pada penelitian ini, dilakukan desain dan rancang bangun
sistem pengenal biometris berbasis citra pembuluh vena pada
telapak tangan. Untuk keperluan pengenalan serta pencitraan yang
baik, dipilih panjang gelombang yang cocok untuk pencahayaan
berintensitas rendah dan penggunaan metode Local Binary Pattern
(LBP) untuk mendapatkan pola pembuluh darah pada telapak
tangan. Citra yang diperoleh dengan LED berpanjang gelombang
740nm dan 810nm menunjukkan gradient kontras pembuluh darah
yang kecil dan masih terdapat pola garis telapak tangan. Maka
Panjang gelombang 740nm dan 810nm dianggap kurang cocok
untuk pendeteksian pembuluh darah pada telapak tangan. Dengan
menggunakan panjang gelombang 940nm, pola terlihat jelas, dan
pola telapak tangan sebagian besar hilang.
Selanjutnya pada pre-processing dilakukan proses
smoothing dan peregangan kontras. Segmentasi citra dilakukan
dengan mendapatkan area ROI untuk mendapatkan informasi
fiturnya. Proses identifikasi fitur citra diperoleh dengan
menggunakan metode Mean Square Error (MSE) dan Local Binary
Pattern (LBP). Pengujian dilakukan dengan 8 sampel yang
dijadikan database sebanyak 1418 kali dan 314 kali untuk 6 sampel
yang tidak termasuk database . Parameter MSE didapatkan sebesar
0,13 sebagai batas kecocokan minimum. Hasil akhir didapatkan
nilai recognition rate total sebesar 90,69% dan error rate sebesar
1,97%.
==========================================================================================================
In this study, conducted the design of biometric identification
system based on the image of the palm veins. For the purposes of
recognition and good imaging, selected appropriate wavelengths
suited for low-intensity lighting and use of Local Binary Pattern
(LBP) methods to obtain the pattern of veins in the palm of the hand.
Imagery obtained by the LED wavelength 740nm and 810nm
showed the contrast gradient of small blood vessels and there is still
a line pattern of palms. Therefore wavelength of 740nm and 810nm
are considered less suitable for the detection of blood vessels in the
palm of the hand. By using a wavelength of 940nm, the pattern is
clearly visible, and the line pattern of the palms is mostly gone.
Furthermore, the pre-processing performed smoothing
process and contrast stretching. Image segmentation is done by
getting the ROI area to get its feature information. The
identification process image features obtained using Mean Square
Error (MSE) and the Local Binary Pattern (LBP) methods. Testing
is done with 8 samples is used as the database with data obtained as
much as 1418 times and 314 times for 6 samples that did not include
a database. The MSE parameters obtained value is 0.13 as the limit
of minimum match. The final result obtained recognition rate total
value amounted to 90.69% and the error rate of 1.97%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSF 621.362 Sya r |
| Uncontrolled Keywords: | ROI, LBP, MSE,Recognition rate,Error rate |
| Subjects: | Q Science > QC Physics > QC457 Infrared technology. |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | - Taufiq Rahmanu |
| Date Deposited: | 26 Nov 2019 06:12 |
| Last Modified: | 26 Nov 2019 06:12 |
| URI: | http://repository.its.ac.id/id/eprint/72039 |
Actions (login required)
 |
View Item |