人脸特征潜在分布规律的人脸注释算法

An algorithm based on the potential distribution of face characteristics for extracted human faces

提出了一种简单有效的人脸注释改进算法。该算法将二值边缘图像的梯度映射和人的面部特征的潜在分布规律相结合。首先应用二值边缘图像的按水平方向以及按垂直方向的梯度的映射,快速确定人脸位置以及人眼的粗略位置;然后对人眼的位置使用逐步逼近法进行进一步的精确定位,再利用图像垂直方向的梯度映射确定嘴的位置;结合人眼和嘴的位置,以及人脸面部特征的潜在分布规律,确定鼻子的位置。

A valid algorithm for extracted human faces and facial features has been put forward. This algorithm has combined the gradient mapping of binary edge image and the potential distribution of people's facial characteristics. First of all, using the gradient mapping of horizontal direction and vertical direction of binary edge image, to determine the location of the human face and the rough location of the human eye, then using the approximation method to location the human eye’s accurate positioning; to determine the location of mouth by the gradient mapping in vertical direction of the image; combination of the human eye position, mouth position, as well as the potential distribution of face characteristics to determine the location of the nose. At the same time, in order to verify the effectiveness of the algorithm, a series of different people in different postures has been used to validate the algorithm.