Nuraini, Ulfa Siti (2019) Identifikasi Wajah Menggunakan Metode Multilayer Perceptron (MLP) dan Convolutional Neural Network (CNN). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211540000021-Undergraduate_Theses.pdf Download (2MB) | Preview |
Abstract
Presensi merupakan sebuah kegiatan pengambilan data guna mengetahui jumlah kehadiran dalam suatu kondisi. Presensi secara manual rentan dalam terjadinya kecurangan misal pemalsuan tanda tangan. Bahkan presensi yang menggunakan teknologi yaitu GPS juga belum bisa menjadi tolak ukur kedisiplinan waktu. Sehingga perlu digunakan sistem biometrik yaitu wajah. Wajah termasuk dalam biometrik fisiologis dengan tingkat tantangan yang lebih besar daripada biometrik fisiologis yang lain. Dalam pengolahan citra, sering menggunakan algoritma machine learning. Metode yang umum digunakan yaitu Neural Network dimana yang memiliki tambahan layer disebut Multilayer Perceptron (MLP). Selain itu teknik lain yang digunakan dalam pengolahan citra yaitu Convolutional Neural Network (CNN). Dalam tugas akhir ini, dilakukan perbandingan metode MLP dan CNN dalam pengolahan citra dengan studi kasus identifikasi wajah. Penelitian ini menggunakan 4 ×10 × 8 citra sebagai input dan variabel target adalah identitas dari 4 orang. Kesimpulan yang didapatkan yaitu metode Convolutional Neural Network yang dilihat dari accuracy, precision, sensitivity, dan Fscore memiliki hasil yang lebih baik daripada metode Multilayer Perceptron.
=================================================================================================================================
Presence is an activity of taking data to find out the number of attendance. Manually presence is susceptible to fraud, like signature forgery. Even presence that uses GPS technology cannot be a benchmark for discipline of time. So, it is necessary used biometric systems, namely face. Face is included in physiological biometrics with a greater level of challenge than other physiological biometrics. Algorithm that is often used for image processing is machine learning. The commonly used method is the Neural Network where the additional layer is called the Multilayer Perceptron (MLP). Besides, other technique used in image processing is Convolutional Neural Network (CNN). In this final project, a comparison of MLP and CNN methods was carried out in image processing with a face identification case study. This study uses 4 × 10 × 8 images as an input and identity of 4 people as a target. The conclusion is the Convolutional Neural Network method which is seen from accuracy, precision, sensitivity, and Fscore, has better results than the Multilayer Perceptron method.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.233 Nur i-1 2019 |
| Uncontrolled Keywords: | Citra, Convolutional Neural Network, Kinerja Klasifikasi, Multilayer Perceptron, Wajah |
| Subjects: | Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Ulfa Siti Nuraini |
| Date Deposited: | 13 Mar 2024 06:06 |
| Last Modified: | 13 Mar 2024 06:06 |
| URI: | http://repository.its.ac.id/id/eprint/64262 |
Actions (login required)
 |
View Item |