摘要
提出了一种新的小波包分解算法,该算法只对图像的低频部分进行分解,从而大大减少了特征向量的维数,提高了运行效率。在模式分类上采用了支持向量机的算法,实验证明,该算法与Fisher线性判别算法相比,其检测率明显提高。
A new algorithm of decomposition has been introduced, namely to decompose the lowpass subbands of image, so the dimensions of characters vector accordingly are decreased largely and the running efficiency is increased. In the pattern discrimination, the support vector machine algorithm is used, the experiments results show, the detection rate is increased evidently than Fisher linear discrimination.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2005年第1期60-62,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60073052)
武警部队军事科研项目
关键词
隐秘检测
小波包分解
特征向量
支持向量机
Detection of steganography
Wavelet packet decomposition
Characters vector
Support vector machines