Abstract: RBF support vector machine (SVM) is widely used in pattern recognition and fault detection because of its stability and high recognition rate. Unlike other literatures, this paper applies RBF support vector machine to multi-class mixing face recognition. On the one hand, after multi-class mixing, the performance of RBF kernel function for support vector machine is examined whether investigated or not. On the other hand, this method has more practical value. Experiments show that the performance of multi-class mixing samples is slightly degraded compared with that of single-class RBF support vector machine, but the recognition rate is still very high, which verifies the effectiveness and practicability of the method.

Key words: Support vector machine, RBF kernel function, multi-class mixing, face recognition

摘要： RBF支持向量机以其稳定性和识别率高的优势被广泛地应用于模式识别和故障检测。与其他文献不同，RBF支持向量机被应用于多类混叠的人脸识别。一方面考察经过多类混叠后，以RBF为核函数的支持向量机的性能有无退化，另一方面使其更具实用价值。经过实验验证，与RBF支持向量机用于单一类别相比，多类混叠的样本在性能上的确稍有退化，但是仍旧保持了很高的识别率，验证了该方法的有效性和实用性。

关键词: 支持向量机, RBF核函数, 多类混叠, 人脸识别