基于FLM的特征检测算法在手势识别中的应用

An arithmetic of feature detection and application based on FLM

提出一种用于特征检测的基于高斯函数的特征相似度函数 (FeatureLikelihoodmap(FLM) ) ,该函数是归一化为[0 ,1]的一个函数 ,它模拟了在尺度空间中各点的图像特征的相似度 .FLM继承了一些基于特征的图像描述方法的特点 ,避免了模型和数据匹配算法中所需要的大量的搜索 ,它还淘汰了基于特征的传统方法中的阈值需要 .本文推导了对称圆状特征的相似度函数 ,并且将它运用于手势识别中 ,实验结果表明该方法能很好地检测出手形特征并且进行手区域定位和手形识别 .

This paper introduce a new approach for feature detection using a feature likelihood map(FLM),which is a function based on Gaussian kernel and normalized to the interval [0,1], and which approximates the likelihood of image features at all points in scale-space. On one hand, FLM inherits several advantages of feature based image representations, on the other hand, it avoids the need for explicit search when matching features in object models to image data, and it also eliminates the need for thresholds present in most traditional feature based approach. We derived the feature likelihood map for symmetric blob-like image structures and analyzed its behavior on synthetic and real images. In the end, we used this method to gesture recognition and the result shows this arithmetic can recognize gestures effectively.