CAJ | 학술논문

针对传统Adaboost人脸检测算法中光照不均匀、图像检测正确率低的问题，提出一种基于Adaboost算法的人脸光照补偿检测方法。介绍Adaboost人脸检测算法的处理流程以及直方图均衡化原理，并将Adaboost算法和直方图均衡化相结合，实现人脸检测。检测结果表明，与传统的Adaboost人脸检测方法相比，新方法对于光照不均匀图像的人脸检测有很好的效果。
침대전통Adaboost인검검측산법중광조불균균、도상검측정학솔저적문제，제출일충기우Adaboost산법적인검광조보상검측방법。개소Adaboost인검검측산법적처리류정이급직방도균형화원리，병장Adaboost산법화직방도균형화상결합，실현인검검측。검측결과표명，여전통적Adaboost인검검측방법상비，신방법대우광조불균균도상적인검검측유흔호적효과。
Traditional Adaboost face detection algorithm uneven illumination image detection rate decreased problem ,we propose a face detection method based on Adaboost algorithm for illumination compensation .Details of the Adaboost face detection algorithm processing and histogram equalization principle and Adaboost algorithm and histogram equalization combined to achieve face detection .Test results show that this method compared with conventional Adaboost face detec-tion method to detect the speed difference is not big ,but uneven illumination image face detection accuracy and false detec-tion rate has very good results .