A Novel and Lightweight Real-Time Continuous Motion Gesture Recognition Algorithm for Smartphones

Advancement in smartphones has facilitated the investigation of new modalities human-machine interaction, including communication through touch, voice, and gestures. In-depth, researchers examined problem recognizing distinct gestures (surface, hand, motion). However, gesture recognition algorithm pitches discontinuity while user performs subsequent continuous gesture. The may occur due to selection a delimiter differentiate between successive motions or employment complex boost accuracy detection, which takes significant time recognize before enter next Further, based on template matching, machine learning models, neural networks requires lot storage space, processing resources, both, are resource-intensive for smartphones. This research proposes novel Axis-Point Continuous Motion Gesture (APCMG) that uses accelerometer sensor data recognizes motion real time. low computational complexity easily implemented resource-constrained devices with minimal computing cost, memory, energy. prime objective APCMG is find start end from stream real-time. To demonstrate efficacy, experimental simulation Android application dialing phone number considered. App acknowledges 12 corresponding 0 9 number, delete, calls termination. simulations collected 7500 samples 25 volunteers. efficiently isolated 95% 94% accuracy, respectively. proposed energy consumption.