Intelligent design and optimization of exercise equipment based on fusion algorithm of YOLOv5-ResNet 50

In the context of a low-carbon economy, it is crucial to integrate environmentally friendly and intelligent designs into fitness facilities to ensure sustainable economic, environmental, and energy development. However, traditional fitness facilities face challenges in effectively monitoring user movements, conserving energy, and optimizing efficiency. The need to conserve energy and reduce resource costs in a low-carbon economy calls for the development of efficient algorithms that can maintain accuracy while minimizing computation and energy consumption. This study proposes a method that combines machine learning and computer vision techniques to enhance monitoring accuracy while minimizing computation and energy consumption. The ResNet-50 model is utilized to extract image features associated with human movements, while real-time object detection and tracking are performed using the YOLOv5 model. Experimental evaluations are conducted on a dataset comprising multiple action categories, and the results demonstrate the excellent performance and model efficiency of the proposed method. Specifically, on the KU-HAR dataset, the proposed algorithm achieves a reduction of over 46.8% in inference time and more than 45.9% in FLOPs, while improving the MAPE by more than 42.8%. These advancements significantly enhance the accuracy and robustness of human motion recognition, highlighting the importance of this approach in the green transformation and intelligent design of fitness facilities. • The proposed approach combines various deep learning algorithms, for better recognition of action. • Focuses on effectively monitoring user movements, conserving energy, and improving equipment efficiency. • Experimental evaluations demonstrate the exceptional performance and model efficiency.