基于CNN的手势游戏控制技术与开发实现

广东东软学院本科毕业设计（论文） 基 于 CNN 的 手 势 游 戏 控 制 技 术 与 开 发 实 现Control technology and development of gesture game based on CNN AbstractHuman-computer interaction is a hot topic in the field of computers. The development of computers cannot only focus on hardware improvement, but also is particularly important in terms of human-computer interaction. Today's interaction technology is also changing with each passing day. Gesture recognition technology as a relatively new interactive technology, it can help people more easily control the software, speed up the efficiency of document processing. Gesture recognition technology can be divided into gesture recognition based on data gloves and other hardware and gesture recognition based on computer vision according to different input devices. There are many types of gestures, which makes it difficult for computers to perform gesture recognition and classification operations. With the improvement of hardware performance in the computer field and more suitable software algorithms, many problems that were impossible to solve before became possible.Gesture recognition in the industry is roughly divided into static gesture 广东东软学院本科毕业设计（论文）recognition and dynamic gesture recognition. The first is that the computer only confirms the user's instructions by recognizing the static image of the hand. The second is that the computer can accurately determine user instructions by recognizing the shape of the hand and recognizing the movement trajectory of the hand. For static gestures, this project proposes an effective CNN-based extraction method. The camera is obtained through the opencv library to obtain the contour of the hand, and finally, the ...