【摘要】 在传统DCV的基础上,提出了一种改进的快速DCV分类方法。该方法与传统的DCV分类方法相比,在保证识别率相同的情况下具有较快的分类速率。传统的DCV分类方法通过计算特征向量之间的距离来进行分类,而所提快速DCV分类方法则通过标量计算完成分类。理论分析及复杂度计算表明,快速DCV分类方法的分类速率是传统DCV分类方法的2倍左右,在Yale、ORL和PIE 3种人脸数据库得到的对比仿真实验结果验证了该算法的有效性。更多还原

【Abstract】 This paper proposed an improved Fast Discriminative Common Vectors(FDCV)classification algorithm based on the traditional Discriminative Common Vectors(DCV).Compared with the traditional DCV,the FDCV not only has a faster classification rate,but also guarantees the same recognition performance.The FDCV does the classification by calculating the distance between scalars but not vectors which are used in the traditional DCV.Theoretical analysis and complexity calculation show that Faster DCV has twice the classifying speed of the traditional DCV.The simulation experiment on the three face databases of Yale Face Database B,ORL Database and PIE Database further verifies the effectiveness of the algorithm.更多还原