International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 11 | Nov 2025
p-ISSN: 2395-0072
www.irjet.net
WaveCanvas Ramesh Sattigeri1 Nandita P2, Nayan M3, Nandini B4, Shivanand K5 1Assistant Professor, SG Balekundri Institute of Technology, Belagavi, Karnataka, India 2,3,4,5Student, SG Balekundri Institute of Technology, Belagavi, Karnataka, India
----------------------------------------------------------------------------------***-----------------------------------------------------------------------------
Abstract - WaveCanvas is an interactive, AI-powered virtual drawing system that enables users to draw in the air using
only hand gestures, without requiring a physical pen, touch screen, or stylus. The system leverages computer vision, deep learning, and gesture recognition to track hand movement in real time. By integrating MediaPipe for hand- landmark detection, Convolutional Neural Networks (CNNs) for gesture classification, and a Flask-based web API for communication, the solution provides a smooth, responsive, and hardware-free drawing experience. The frontend—built with HTML, CSS, and JavaScript—renders live strokes on a digital canvas. The project uses self-prepared datasets for training gesture models, ensuring high accuracy across different lighting and usage conditions.This paper presents the complete methodology, system architecture, software pipeline, training dataset preparation, performance analysis, and advantages of the AirCanvas system. Keywords: WaveCanvas, Gesture Recognition, MediaPipe, CNN, Machine Learning, Flask API, Virtual Drawing, HumanComputer Interaction.
1. INTRODUCTION Human–computer interaction is shifting rapidly from device- dependent controls to natural, touchless interfaces. Traditional drawing tools—such as stylus pens, mouse devices, or touch panels—restrict mobility and require physical contact. As creative industries and educational platforms move towards immersive digital experiences, the need for contactless, intuitive drawing systems becomes essential. Aircavas bridges this gap by allowing users to “draw in the air” using only their index finger. The system captures hand movement through a webcam, processes gestures using MediaPipe and CNNbased models, and translates finger trajectories into real-time strokes on a digital board.Inspired by modern gesture-based systems but designed with simplicity and accessibility, the AirCanvas platform aims to provide:
Contactless drawing Real-time gesture tracking ML-based gesture recognition A lightweight, user-friendly web interface Simple integration via Flask APIs
Research Objectives The primary objectives of this research project are: To design a contactless drawing interface using hand gestures. To implement MediaPipe-based hand tracking for precise finger-tip detection. To train a CNN model on custom gesture datasets for mode switching (draw, erase, select color, clear screen, etc.). To build a backend using Flask API for efficient communication between machine-learning modules and the web interface. To create a responsive frontend using HTML, CSS, and JavaScript for displaying the AirCanvas output. To ensure real-time performance with minimal latency and high accuracy in gesture classification.
1.2 Operational Modes
User-Facing Features Draw using hand movements Select colors via gestures Erase specific parts Clear complete canvas Control brush thickness Smooth, real-time output on browser
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 867