The existing face detection methods can be classified into four categories: Knowledge-based methods, Feature invariant approaches, Template matching methods and Appearance-based methods.
现有的静止图像人脸检测算法主要有四类:基于知识的方法、基于特征不变量的方法、基于模板匹配的方法和基于表象的方法。
In essence it takes the decision profile and decision template as target's feature vector and template respectively, and adopts template matching method to target recognition.
该方法实质上是将目标的决策分布图作为特征矢量,将其和每类目标的决策模板进行匹配以实现融合识别。
In the detection of motion objects, a method based on movement feature of target (catenary) is proposed. It combines temporal differential method and template matching in image sequence data.
在检测运动物体方面,根据检测对象(接触线)的运动特征提出了一种对图像序列数据采用时态差分法与模板匹配相结合的检测方法。
应用推荐