【摘要】 为了解决人们长时间用同一个姿势刷小视频或者久坐不动而引起的健康问题，本文提出了一种基于卷积神经网络和OpenCV库相结合的手势模仿小游戏。游戏以预训练的ResNet-50模型作为基础网络架构，利用PyTorch框架作为深度学习的基本框架，并利用卷积神经网络算法来构建手势识别模型。用户可通过该小游戏来不断模仿目标手势，以满足日常锻炼的需要。更多还原

【Abstract】 In order to solve the health problems caused by people using the same posture to brush small videos for a long time or sitting still for a long time, this article proposes a gesture imitation mini game based on a combination of convolutional neural networks and OpenCV library. The game uses the pre trained ResNet-50 model as the basic network architecture, the PyTorch framework as the basic framework for deep learning, and convolutional neural network algorithms to construct gesture recognition models. Users can continuously imitate target gestures through this mini game to meet their daily exercise needs.更多还原