Virtual mouse design based on OpenCV and deep learning

Gesture control technology enables non-contact operation of electronic devices through advanced sensors and algorithms. This design employs deep learning for gesture recognition, using Python and OpenCV to call the camera and process images, expanding the dataset to 4590 images through data augmentation. The VGG neural network model is built and trained using Keras, which reads camera frames in real-time to predict gestures and simulate mouse operations. This virtual mouse simplifies human-computer interaction, enhances user experience, and addresses the difficulties faced by the elderly, people with disabilities, and those in educational or presentation settings who find traditional mouse usage inconvenient.