重庆理工大学学报：自然科学
重慶理工大學學報：自然科學
중경리공대학학보：자연과학
Journal of Chongqing Institute of Technology
2012年
3期
104-108
,共5页
人脸检测%AdaBoost算法%分类器%支持向量机%模式识别
人臉檢測%AdaBoost算法%分類器%支持嚮量機%模式識彆
인검검측%AdaBoost산법%분류기%지지향량궤%모식식별
face detection%AdaBoost algorithm%classifiers%support vector machine%pattern recog-nition
用支持向量机建立新的核函数，使得该函数集集成无穷多个AdaBoost算法的弱分类器，最终形成强分类器。应用该强分类器进行人脸检测。实验结果表明，该方法的人脸检测率优于有限维AdaBoost算法，提高了检测精度。
用支持嚮量機建立新的覈函數，使得該函數集集成無窮多箇AdaBoost算法的弱分類器，最終形成彊分類器。應用該彊分類器進行人臉檢測。實驗結果錶明，該方法的人臉檢測率優于有限維AdaBoost算法，提高瞭檢測精度。
용지지향량궤건립신적핵함수，사득해함수집집성무궁다개AdaBoost산법적약분류기，최종형성강분류기。응용해강분류기진행인검검측。실험결과표명，해방법적인검검측솔우우유한유AdaBoost산법，제고료검측정도。
Face detection is a basic and important research subject in the machine vision and pattern recognition, which has important application value in image, video retrieval, video monitoring, automatic face recognition and intelligent human-machine interaction etc. After long-term development, face detection method has made remarkable achievements, and is widely used in many of the research method. Based on support vector machine, a new kernel function was built, which contains infinite weak classifiers of AdaBoost algorithm, and those weak classifiers finally formed a strong classifier. The experimental results show that face detection rate based on this method is better than the limited AdaBoost algorithm, which improved the detection accuracy.
