Text

PROTOTIPE SISTEM KEAMANAN AKSES MASUK RUANGAN BERBASIS VOICE RECOGNITION MENGGUNAKAN ALGORITMA DEEP LEARNING

---

Penilaian

---

The security system that is often used today is the conventional security system. This security system has weaknesses that allow it to be imitated and broken into. Therefore, a security system that uses biometrics is needed. The human voice is one form of biometrics that can be used to identify people (person identification) which can be utilized in the security field. The purpose of this research is to implement the use of voice as a room access security system by using deep learning algorithms in the learning process to recognize voices in real-time. In this research, there are several stages carried out, namely taking voice data, processing voice data, training voice data using 3 CNN architectures, offline testing using testing data and online testing using microphones in real time. The voice data processing used is Short-Time Fourier Transform (STFT) for the feature extraction process in the form of spectrograms. Then, the spectrogram will be processed by Convolutional Neural Network (CNN) as an identifier method. This research uses 3 types of CNN architecture, namely VGG16 architecture, AlexNet, and Own Model Architecture and obtained the best accuracy results from the 3 architectures with epoch 100 of 0.9993, 0.9915, and 1. Testing is done offline and online, offline testing is done using test data and using 3 architectures, obtained the best test results using epoch 100 of 61.76%, 94.11% and 95.09%. Furthermore, online testing is carried out in real time on a prototype in the Basic Laboratory of Control Systems and Robotics using a microphone against 15 respondents consisting of 8 respondents who are in the dataset and 7 who are not in the dataset. The real time test results obtained an average voice recognition accuracy value of 80%. The results show that the model of its own architecture has better performance for recognizing voices, and can be implemented in the room access security system.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2307002797	T122663	T1226632023	Central Library (Referens)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1226632023

Publisher

Inderalaya : Jurusan Teknik Elektro, Fakultas Tenik., 2023

Collation

xiv, 90 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

621.990 7

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Jurusan Teknik Elektro
safety equipment/perkakas keamanan

Specific Detail Info

-

Statement of Responsibility

SEPTA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment