IMPLEMENTASI PENGENALAN OBJEK DAN WAJAH PADA SERVICE ROBOT BERBASIS CONVOLUTIONAL NEURAL NETWORK

SEPTYANTO, AHMAD and Irmawan, Irmawan (2024) IMPLEMENTASI PENGENALAN OBJEK DAN WAJAH PADA SERVICE ROBOT BERBASIS CONVOLUTIONAL NEURAL NETWORK. Undergraduate thesis, Sriwijaya University.

	Text RAMA_20201_03041382025099.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (9MB) | Request a copy
	Text RAMA_20201_03041382025099_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (7MB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_20201_03041382025099_0017097405_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (630kB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (67kB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (140kB) | Request a copy
	Text RAMA_20201_03041382025099_0017097405_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (803kB) | Request a copy

Abstract

Abstrak—Service robot merupakan sebuah humanoid robot yang dapat melaksanakan tugasnya diberbagai aspek kehidupan seperti pada sektor industri, pendidikan, wisata, serta sosial. Dalam menjalankan fungsinya, service robot perlu memiliki kemampuan untuk mendeteksi dan mengenali objek sekitar. Penelitian sebelumnya menunjukkan metode deteksi objek dan wajah belum diimplementasikan pada robot secara real-time Penelitian ini bertujuan untuk deteksi objek dan wajah secara real-time. Selain itu, pada penelitian ini juga mengembangkan metode estimasi jarak dengan menggunakan informasi dari monocular distance estimation. Serta melihat performansi dari algoritma YOLOv8 dalam melakukan pendeteksian objek dan VGG19 pendeteksian wajah secara real time. Dataset yang digunakan terdiri dari 11 kelas objek dan 30 wajah mahasiswa. Hasil pelatihan model objek Yolov8 mendapatkan tingkat keberhasilan pengenalan objek sebesar 80,7% dengan loss terendah 0,45121. Sedangkan model wajah menggunakan VGG19 adalah 80%. dengan loss 0.00665. Hasil dari pengukuran jarak objek maupun wajah yang dikenali menggunakan metode monocular distance estimation sudah sangat baik dengan mendapatkan Mean Avarage Error yang didapatkan 1.525 cm untuk model pengenalan objek dan 0.812 cm untuk model pengenalan wajah. Penelitian ini menunjukkan bahwa implementasi dan pengembangan Convolutional Neural Network dengan arsitektur Yolov8 dan model VGG19 untuk pengenalan objek dan wajah pada service robot telah berhasil dilakukan dengan baik, seperti terlihat dari tingkat akurasi yang mencapai hasil yang baik untuk pengenalan objek dan wajah.dan manusia, memungkinkan robot memberikan respon yang sesuai terhadap input suara yang diterima. Kata kunci: Service robot, Convolutional Neural Network, YOLOv8, VGG19, Monocular Distance Estimation Abstract—Service robot is a humanoid robot that can carry out its duties in various aspects of life such as in the industrial, educational, tourism, and social sectors. In carrying out its functions, service robots need to have the ability to detect and recognize surrounding objects. Previous research shows that object and face detection methods have not been implemented on robots in real-time This research aims to detect objects and faces in real-time. In addition, this research also develops a distance estimation method using information from monocular distance estimation. And see the performance of the YOLOv8 algorithm in object detection and VGG19 face detection in real time. The dataset used consists of 11 object classes and 30 student faces. The results of Yolov8 object model training get an object recognition success rate of 80.7% with the lowest loss of 0.45121. While the face model using VGG19 is 80%. with a loss of 0.00665. The results of measuring the distance of objects and faces recognized using the monocular distance estimation method are very good by getting the Mean Avarage Error obtained 1.525 cm for the object recognition model and 0.812 cm for the face recognition model. This research shows that the implementation and development of Convolutional Neural Network with Yolov8 architecture and VGG19 model for object and face recognition on service robot has been successfully done well, as seen from the accuracy level that achieves good results for object and face recognition. Keywords: Service robot, Convolutional Neural Network, YOLOv8, VGG19, Monocular Distance Estimation

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Service robot, Convolutional Neural Network, YOLOv8, VGG19, Monocular Distance Estimation
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1-9971 Electrical engineering. Electronics. Nuclear engineering > TK1 Electrical engineering--Periodicals. Automatic control--Periodicals. Computer science--Periodicals. Information technology--Periodicals. Automatic control. Computer science. Electrical engineering. Information technology.
Divisions:	03-Faculty of Engineering > 20201-Electrical Engineering (S1)
Depositing User:	Ahmad Septyanto
Date Deposited:	18 Jul 2024 02:07
Last Modified:	18 Jul 2024 02:07
URI:	http://repository.unsri.ac.id/id/eprint/151446

Actions (login required)

View Item