Unique Feature Extraction and Consistency Network for Skeleton Body Keypoints Configuration and Enhancement

The machine learning algorithm recognizes skeleton poses in many ways but previous methods suffered from higher computation and low latency. To increase solitary person posture accurateness, we developed a novel strategy for fitness-related appliances. The developed design included a basic network that offered a starting point for eminent modification using a Unique Features Extraction Network (UFE Net) for long-term consistency. The UFE Net imposed innovative constraints through local element changes and enhancement of global body properties. The recommended module decreased the effect of varying body key points and inconsistent postures. Using a standard dataset, we showed how effective the suggested approach was. The proposed framework operated rapidly and increased precision on the CPU environment platform. In comparison to previous results, the developed design enhanced accuracy by 98.5% at real-time proceeding rates of 32 FPS.