Dianta, Ashafidz Fauzan (2015) Pengenalan Seseorang Berbasis Skoring Data Trayektori Gaya Berjalan (GAIT) Menggunakan Naive Bayesian. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2213205202-Master_Theses.pdf - Accepted Version Download (2MB) | Preview |
Preview |
Text
2213205202-Paper.pdf - Accepted Version Download (397kB) | Preview |
Preview |
Text
2213205202-Presentation.pdf - Presentation Download (707kB) | Preview |
Abstract
Gaya berjalan seseorang adalah sangat unik, sebab cara berjalan seseorang sangat tergantung pada anggota tubuh bagian bawah seseorang seperti struktur tulang, otot, ligament dan tendon orang tersebut, hal itulah yang menjadikan gaya berjalan seseorang menjadi unik. Penelitian ini bermaksud, dapatkah kita mengenali seseorang dari gaya berjalannya. Data lintasan marker / marker trajectory digunakan sebagai input sistem biometrik. Dalam penelitian ini, diukur 8 orang sehat dan subjek tersebut diarahkan untuk berjalan secara normal di lintasan berjalan di laboratorium motion capture, dengan kecepatan jalan sesuai kenyamanan masing-masing subjek. Pada kaki kanan dan kiri, subjek tersebut ditempelkan marker sebanyak 16 buah (lower limbs). Setiap marker akan memiliki data lintasan 3D, yaitu lintasan pada sumbu X, sumbu Y dan sumbu Z. Data marker trayektori diskoringkan, dan Naive Bayesian Algorithm digunakan untuk mengenali seseorang melalui gaya berjalannya. Jumlah keseluruhan data marker yang dianalisis sebanyak 50 data skoring marker trayektori gaya berjalan. Setiap satu data skoring trayektori marker gaya berjalan terdapat data trayektori marker selama satu gait cycle.
Untuk uji coba, pada penelitian ini digunakan 16 data trayektori marker gaya berjalan untuk dijadikan data training ke dalam sistem pengenalan, dan sebanyak 34 data trayektori marker gaya berjalan digunakan sebagai data uji sistem biometrik. Dari hasil uji coba tersebut, 98,24% data uji dapat dikenali oleh sistem siapa pemilik gaya berjalan tersebut, dan 1,76% sistem tidak tepat dalam mengenali pemilik gaya berjalan tersebut, dengan hasil tersebut dapat disimpulkan bahwa data marker trayektori dapat dijadikan salah satu alternatif untuk proses mengenali sistem biometrik seseorang. ========== Human gait is unique and specific to every one, this is because human gait depends fully on the structure of the human lower limbs, such as bones, muscles, ligaments and tendons. This study intends to recognize of a person in gait motion, and can it be used as a biometric of someone. Marker trajectory data was used as an input in this biometric systems. In this study, eight healthy subjects were asked to walk with self-selected speed in the gait lab equiped with motion capture system. Sixteen reflective markers were attached to the bont landmarks of lower limbs. Marker trajectories data in 3D were then obtained, trajectory in X, Y and Z plane. One full gait cycle was used as one input data in the system. Scoring algorithm based on the trajectories pattern was used to score each trajectory data before being conveyed to the Naive Bayesian Classifier. Since every subject walked for three times, and within one trial walking there could be about 2-3 gait cycle of one full gait cycle, in total 50 marker trajectories data of one gait cycle from 8 healthy subjects were used for biomectrics simulation.
In this experiments, 16 gait cycle data marker trajectory were used as a training marker data, and 34 others were used as testing data marker trajectory. The result showed that 97% of testing data can be recognized by the system biometrics, and 3% of the testing data were falsely recognized by the system. In conclusion, recognition sistem based on the human gait pattern has proven the unique of human walking and can be used as a biometrics system.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTE 519.542 Dia p 3100015062714 |
| Uncontrolled Keywords: | Marker Base Biometric, Gait Analysis, System Biometric Gait, Marker Trajectory Data, Marker Base Biometrics, Gait Analysis, System Biometric Gait, Marker Trajectory Data |
| Subjects: | Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory. Q Science > QP Physiology > QP310.W3 Gait in humans. |
| Depositing User: | - Davi Wah |
| Date Deposited: | 06 Dec 2019 07:22 |
| Last Modified: | 06 Dec 2019 07:22 |
| URI: | http://repository.its.ac.id/id/eprint/72207 |
Actions (login required)
 |
View Item |