项目名称: 基于MEMS加速度传感器的智能终端手势识别及三维交互模型
项目编号: No.61501207
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 无线电电子学、电信技术
项目作者: 何振宇
作者单位: 暨南大学
项目金额: 20万元
中文摘要: 基于加速度传感器的智能终端手势识别及人机交互,是通过无处不在的智能终端上的加速度感器采集用户手势动作数据,然后利用手势识别算法,实现更为自然的人机交互方式。手势识别是一种重要的人机交互技术,目前通常采用基于图像处理的方法,由于该方法对光线、背景等外部条件的依赖比较强,对特定的场合和人群的应用时,存在局限性。而本项目采用加速度传感器的方法正好可以克服这些弱点。本项目的具体研究内容包括:(1)采用ADXL330三轴加速度传感器,建立用户手势加速度信号的标准数据库;(2)深入研究加速度手势信号的特征提取方法和手势轨迹恢复;(3)面对资源受限的终端设备,优化分类器,提高算法速度和降低内存占用。本项目完成的意义在于为以后的人机交互的发展探索出一条新的途径。
中文关键词: 手势识别;加速度传感器;智能终端;特征提取;轨迹恢复
英文摘要: Gesture recognition and human-computer interaction (HCI) based on the Accelerometer of intelligent terminal, which attempts to collect user’s gesture acceleration signals using the accelerometer of intelligent terminal and then carry out gesture recognition algorithm to achieve a more natural HCI. Gesture recognition is an important human-computer interaction technique. At present, most gesture recognition base on image processing method, which has a lot of limitations. For example, it dependent on external conditions such as light, background and so on. Fortunately, our method can overcome these weaknesses. The research includes following aspects: (1).We establish a standard database of gestures acceleration signal based on ADXL330 tri-axial accelerometer; (2).According to the characteristics of accelerometer signals, we explore different efficient features extraction methods and reconstruct gesture trajectory; (3) Since the resources of the terminal equipment is limit, we optimize the classifier through improving the algorithm speed and reducing memory costs .The significance of the completion of the project is to explore a new way for the future development of HCI.
英文关键词: Gesture recognition;Accelerometer;Intelligent Terminal;Feature extraction;Trajectory Reconstruction