A Lightweight Network Deployed on ARM Device for Hand Gesture Recognition

Enhancing interactions with natural and user-friendly computer interfaces (HCIs) requires the use of hand gestures. This article illustrates a particularly created brilliant neural network that can identify the flat-based platform's movements. Real-time applications on low-power devices can benefit from the suggested approach, which is solely software-implemented and tailored for resource-constrained contexts. To classify hand gestures with high accuracy and minimal computing complexity, the network analyzes input from infrared (IR) sensors. Particular care is taken to reduce the model's memory and processing needs without sacrificing recognition efficiency. The system is tested using a variety of hand gestures and shows that, despite having little hardware resources, it is capable of quickly and accurately recognizing movements. The software solution offers a scalable method for integrating gesture detection systems into a range of HCI applications on wearable and embedded devices. The system demonstrated competitive performance and low power consumption following rigorous testing. This study demonstrates the practical application of lightweight, software-driven machine learning models in fields including home automation, assistive technology, and robotics.