Krishna Raja, Mr. N. V. Murali and Rathnam, S. Swathi and Aishwarya, Ch. Sri Maha Lakshmi and Chandrika, Ch. Geetha and Vyshnavi, M. and Ratnam, T. Swathi and Veekshita, K. Vedha (2025) AI-Based Virtual Mouse. International Journal of Innovative Science and Research Technology, 10 (4): 25apr827. pp. 2880-2885. ISSN 2456-2165

[thumbnail of IJISRT25APR827.pdf] Text
IJISRT25APR827.pdf - Published Version

Download (656kB)

Abstract

An inventive device that enables hands-free computer interaction is the AI-Based Virtual Mouse, which uses generative AI and computer vision. Unlike traditional wireless or Bluetooth mice, which rely on physical components like batteries and dongles, this technology operates without the need for additional hardware. Instead, it recognizes hand gestures using an inbuilt camera or webcam, allowing users to move the cursor, scroll, and left- and right-click without actually touching a mouse. The system was developed in Python and uses OpenCV for computer vision tasks and generative AI models for gesture detection and fingertip tracking. Additionally, MediaPipe and OpenCV provide machine-learning methods for tracking hand movements and translating them into screen cursor movements. The system uses real-time video data to recognize and interpret hand movements to deliver accurate and efficient computer navigation. By eliminating the need for external input devices, our AI-powered virtual mouse enhances convenience and cleanliness while lowering hardware dependencies. Because of its effectiveness on a standard CPU, it is suitable for real-world applications and offers a creative, intuitive, and touch-free alternative to conventional computer interaction methods.

Item Type: Article
Subjects: T Technology > T Technology (General)
Divisions: Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science
Depositing User: Editor IJISRT Publication
Date Deposited: 10 May 2025 10:33
Last Modified: 10 May 2025 10:33
URI: https://eprint.ijisrt.org/id/eprint/808

Actions (login required)

View Item
View Item