CAJ | 학술논문

考虑到人脸表情识别问题在未来的科学应用中可能出现的样本分布不均匀的情况，在提高识别率的基础上，针对这类问题进行了实验研究，将一种改进的AdaBoost算法与SVM结合运用到表情分类当中。实验结果表明，在出现稀有样本的情况下，相对于普通的AdaBoost训练SVM以及单纯的SVM进行多分类的方法，该算法在识别率方面有了很大提高。
고필도인검표정식별문제재미래적과학응용중가능출현적양본분포불균균적정황，재제고식별솔적기출상，침대저류문제진행료실험연구，장일충개진적AdaBoost산법여SVM결합운용도표정분류당중。실험결과표명，재출현희유양본적정황하，상대우보통적AdaBoost훈련SVM이급단순적SVM진행다분류적방법，해산법재식별솔방면유료흔대제고。
Taking into account the uneven sample distribution of facial expression recognition problem tilat may appear in scientific applications in the future, experimental study of such problems is done on the basis of improving the recognition rate, anti it conbines the improved AdaBoost algorithm and SVM, which is applied to the expression classification. The experimental results show that the algorithm has been greatly improved in terms of recognition rate in the case of appearing rare samples, by comparing to ordinary AdaBoost training SVM and simple SVM nmlti-classification method.