New framework for unsupervised universal steganalysis via SRISP-aided outlier detection

作者：

Highlights：

• Eliminating the need for training and avoiding the mess of model mismatch.

• Considering the effect of cover variation on the existing steganalysis features.

• Proposing a new unsupervised universal steganalysis framework via SRISP-aided outlier detection.

• There are two main differences compared to traditional unsupervised outlier detection framework.

• The effectiveness of the proposed framework is proved qualitatively and quantitatively.

摘要

•Eliminating the need for training and avoiding the mess of model mismatch.•Considering the effect of cover variation on the existing steganalysis features.•Proposing a new unsupervised universal steganalysis framework via SRISP-aided outlier detection.•There are two main differences compared to traditional unsupervised outlier detection framework.•The effectiveness of the proposed framework is proved qualitatively and quantitatively.

论文关键词：Steganalysis,Similarity retrieval,Unsupervised outlier detection

论文评审过程：Received 26 January 2016, Revised 19 May 2016, Accepted 19 May 2016, Available online 20 May 2016, Version of Record 6 June 2016.

论文官网地址：https://doi.org/10.1016/j.image.2016.05.011