基于单目跟踪-神经网络分类算法的手势识别

Gesture Recognition based Monocular Tracking using Neural Network

随着人工智能的发展,具有用户界面的空中机器人逐渐普及大众。手势是人类沟通的直观方式,各研究工作都致力于利用自然的手势控制空中机器人。但是,由于该方面的技术还未成熟,尚未解决存在的长久问题。例如从设计角度选择手势的原则,从硬件角度考虑硬件的要求,考虑数据的可用性以及从实际角度考虑算法的复杂性。因此本设计考虑了实际的场景、硬件成本以及算法的适用性,致力于建立一个经济的单目系统,设计简洁直观的手势,将其映射到丰富的目标方向并实现微调。

Hand/arm gestures are an intuitive way to communicate for humans, and various research works have focused on controlling an aerial robot with natural gestures. However, the techniques in this area are still far from mature and many issues in this area have been poorly addressed. For example, there are the principles of choosing gestures from the design point of view, hardware requirements from an economic point of view, consideration of data availability, and algorithm complexity from a practical perspective. The work has jointly considered practical scenarios, hardware cost and algorithm applicability and focused on building an economical monocular system particularly designed for gestures. In addition, natural arm gestures are mapped to rich target directions and convenient fine adjustment is achieved.