手机导航键采用的人机交互方法存在操控不方便、实现较复杂以及成本高等缺点,为此,提出一种结合肤色模型、亮度变换和几何特征的人机交互系统.通过手机前置摄像头采集用户拇指轻微运动的序列图像,在YCbCr颜色空间对图像进行色度判决;基于目标和背景的亮度特征对图像进行亮度变换,并引入值心概念计算前景的几何中心,判断值心在二维指令特征平面中的具体位置.实验结果表明,该人机交互方式能快速方便地识别大拇指发出的4条主指令和12条辅助指令.%The navigation interactive methods for mobile phones have some drawbacks which contain inconvenient operation, difficult implementation and high cost. Aiming at these problems, this paper proposes a novel human-computer interaction system combining with shin color model, intensity transformation and geometrical characteristic. It collects serial images of thumb moving by the front camera of mobile phone, and judges the hue of the image in YCbCr color field. It adjusts the intensity of the image based on object and background gray characteristic, and calculates the geometric centre by the centre of gray value. It finds the relative position of thumb moving in the instruct plane. Experimental results demonstrate that the interaction can recognize four main instructs and twelve auxiliary instructs friendly and quickly.
展开▼
