Virtual Mouse Menggunakan Deep Learning Berbasis Hand Gesture

Hartono, Alfredo Gerald (2023) Virtual Mouse Menggunakan Deep Learning Berbasis Hand Gesture. Other thesis, Institut Teknologi Sepuluh Nopember Surabaya.

Text 07211940000050-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (14MB) | Request a copy

Abstract

Salah satu perangkat utama dalam mengoperasikan sebuah perangkat komputer adalah mouse. Mouse masih memiliki Batasan-batasan dalam penggunaannya, seperti harus digunakan pada alas yang datar, alas tidak terlalu gelap dan tidak terlalu terang, dan lainnya. Sehingga masih sulit mengoperasikan komputer pada saat melakukan kegiatan yang mengharuskan penggunanya untuk berada jauh dari komputer seperti saat presentasi. Oleh sebab itu, pada penelitian virtual mouse yang menggunakan Deep Learning berbasis Hand Gesture ini, akan dilakukan penerapan klasifikasi pose tangan dengan menggunakan Convolutional Neural Network (CNN). Pada penelitian kali ini diperoleh nilai akurasi training berkisar pada angka 99%. Dan pada pengujian model menggunakan metode confusion matrix, diperoleh nilai akurasi 100% pada gambar testing tangan berjarak 50cm dari kamera, 89% pada jarak 100cm dari kamera, serta 48% pada jarak 150cm dari kamera. Sehingga dapat dikatakan bahwa semakin jauh jarak tangan dari kamera, maka semakin sulit bagi sistem virtual mouse untuk melakukan kerja dari fungsi mouse dengan akurat
====================================================================================================================================
One of the main devices in operating a computer device is a mouse. The mouse still has limitations in its use, such as having to use it on a flat surface, the mat is not too dark and not too bright, and so on. So it is still difficult to operate the computer when carrying out activities that require users to be away from the computer such as during presentations. Therefore, in this research, Virtual Mouse Using Hand Gesture-Based Deep Learning, the classification of hand poses will be carried out using a Convolutional Neural Network (CNN). In this study, the training accuracy value was obtained around 99%. And in the testing model using the Confusion Matrix method, an accuracy value of 100% is obtained in testing hand images that are 50cm from the camera, 89% at a distance of 100cm from the camera, and 48% at a distance of 150cm from the camera. So it can be said that the farther the hand is from the camera, the more difficult it is for the virtual mouse system to accurately perform the work of the mouse function

Item Type:	Thesis (Other)
Uncontrolled Keywords:	CNN, Confusion Matrix, Layer, Mouse, Hand gesture
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Alfredo Gerald Hartono
Date Deposited:	25 Aug 2023 02:07
Last Modified:	25 Aug 2023 02:07
URI:	http://repository.its.ac.id/id/eprint/101757

Actions (login required)

View Item