Novel System based on Amalgamation of Embedded systems and Edge AI for Hand Gesture Recognition: An Edge AI Solution for Hand Gesture Recognition
Gargi Singh, Saanika Gautam, Yugnanda Puri, Jolly Parikh, Gargi Mishra · 2024
Hand gesture recognition is highly significant and a natural means of human-computer interaction. This detection is carried out by using the MPU- 6050 sensor, a widely available Inertial Measurement Unit (IMU). MPU-6050 sensor measures the orientation and objects acceleration and is a low-cost, high-performance sensor. NanoEdgeAI studio, a low-power edge computing framework known to process data efficiently, has been employed to implement the accurate machine learning model. In addition to that, the usage of a microprocessor, Nucleo-F401RE, demonstrates the feasibility of deploying robust and responsive hand gesture recognition systems in edge computing scenarios, showcasing the potential for enhanced human-machine interaction. The selected model used for the research is MLP (multilayer perceptron) which is also known for its flexibility and ability. A balanced accuracy a percentage measure of signals correctly identified by the model was found out to be of 84.27%