Abstract: Aiming at the complex computation of existing evaluation methods and weight setting methods are not reasonable enough, this paper proposes a method to evaluate and compare different steganalysis algorithms. The weighted grey correlation degree between the target index vector and optimal index vector is obtained by the grey correlation synthetical evaluation method, and steganalysis algorithms’ performances is sorted by the size of the weighted grey correlation degree. The instance simulation result proves that the proposed method can reasonably evaluate the steganalysis algorithm and verifies it is a feasible method.

Key words: steganalysis, performance evaluation, entropy method, grey relevance degree

摘要： 针对现有隐写分析评估方法计算复杂、权重分配不合理的缺点，结合熵权与灰色关联度对隐写分析算法进行评估。确定算法的主要性能指标，采用熵权法计算各指标的权值，运用灰色关联度综合评价方法计算待评价指标向量与最优指标向量间的加权灰色关联度，并根据关联度的大小对隐写分析算法的优劣进行排序。实例仿真结果表明，该方法能合理评估各类隐写分析算法的优劣，具有可行性。

关键词: 隐写分析, 性能评估, 熵权法, 灰色关联度

CLC Number: 

• TP391