基于分类特征选择的激光成像隐写算法研究

Laser imaging steganalysis algorithm based on classification feature selection research

以往提出的激光成像隐写算法只初步实现了隐身通信,算法的有效性和隐身数据量均需要得到提升。因此,提出基于分类特征选择的激光成像隐写算法。算法使用分类特征选择分析激光成像结果中图像信息来源,对成像过程产生的事物共性特征加以选择,给出分类特征选择样本,构造样本高频数据的正态函数,获取正态函数中的激光成像特征分布因子,给出激光成像隐写算法的写入位置。在算法的具体实现中,将激光成像看成位平面并切割成数据块,把写入位置对应到相应的数据块上直接进行隐写。经实验分析可得,所提算法的有效性和隐身数据量均较以前提出的激光成像隐写算法有很大提升。

Only stealth communication was achieved preliminarily in the laser imaging stealth algorithm, while the effectiveness and invisible data need to be improved. Therefore, the laser steganalysis algorithm based on classification feature selection is put forward. It analyzes the information source of laser imaging by using classification feature selec- tion, select the common feature caused during imaging, give the sample of classification feature, build the sample nor- mal function of high data, obtain the distribution factor of laser imaging feature in normal function, and give the input position of laser imaging steganalysis algorithm. In the implementation of the algorithm, treat laser as a flat and cut it into data blocks and put the input position directly for steganography. The experimental results show that its effective- ness and invisible data are much more improved than before.