Real-Time Hand Gesture Calculator Using Convolution Neural Network

In recent years, the use of hand gestures has become increasingly significant in facilitating seamless human-computer interaction. In this sense, gesture recognition is quite important in our day-to-day lives. We are developing a hand gesture-based calculator that will help dyslexia persons or people having difficulties to comprehend mathematical concepts more readily. There are several applications for gesture recognition, including those in virtual reality, robotics, computer gaming, and sign language. This demonstrates how scientists’ primary research focus should be gesture recognition. We developed a quick and easy method of simple calculation using hand gestures. We created a custom dataset and built a Convolution Neural Network for the training purpose, compared to the existing system we were able to achieve a training accuracy percentage of 98.13%. Our system first identifies the hand segments in the live video feed, then recognizes the gesture using a pre-trained custom dataset, and then assigns numbers in the live feed. The goal of the study is to achieve human-computer connection that is as natural as human-human contact.