基于手机平台的人脸检测系统的研究

随着通信技术和图像处理技术的发展,手机不仅是一种便携的通信工具,而且能够用来拍照。针对市场上的非智能手机拍照的性能欠佳又不具备图像处理功能的缺点,选取了扩展的Haar特征,利用Intel开源的图像处理软件OpenCV在智能手机操作系统上实现了一个Adaboost人脸检测系统,描述了建立样本、图像预处理、特征提取、构造级联分类器和目标检测等过程。多次实验证明,该系统正检率高,检测时间短,具有一定的实用性。

Research of Face Detection System Based on Mobile Phone Platform

With the development of the communication and image processing technology,mobile phones are not only the portable communicating tools but also can be used to practice photography.Aiming at poor performance of non-smart phones photographs and the disadvantages of image processing functions,the extended Haar-Like feature is researched.A face detection system based on Adaboost face detector algorithm is implemented on smartphone OS with the use of OpenCV which is Intel’s open source image processing software.Some processes,such as sample building,image preprocessing,feature extraction,constructing cascade classifier and target detection,are described.Experimental results prove that the system with short response time and high correct detection rate has a certain practicality.