Innovative gesture-based virtual mouse: crafting an ML-driven gesture recognition system for touchless mouse control

Gesture recognition is a popular method for expressing emotions or facilitating communication between people and machines. Hand gesture recognition is a challenging research area because it involves analyzing complex patterns with many variables. This complexity can impact the machine learning models performance. To address this, techniques like feature selection and feature extraction help simplify the data while keeping important information. Traditional approach involved using statistical functions that measure central tendencies and for model optimization, the automatic approach utilized tools like MediaPipe, OpenCV(3.4.2) and TensorFlow. For performance analysis, scikit-learn was used, and visualizations were created with matplotlib. Ultimately, the model delivered an impressive accuracy of 92%.