基于边缘特征的最小化失真视频隐写平台设计

Design of Platform of Minimizing Distortion Video Steganography Based on Edge Features

提出一种基于边缘特征的最小化失真视频隐写方案,并将其实现于嵌入式平台。该设计选用帧内预测隐写模式,对算法做出以下改进:构造基于边缘特征的代价函数,使加密数据集中在难以检测的边缘区域,降低对空间相关性的破坏;在对隐写失真的度量中引入最小化失真框架,解决失真量化不足的问题,提高安全性;编解码过程中引入基于边缘强度的映射规则,提高嵌入效率,减小预测偏差。实验证明,隐写方案在确保视频质量和编码效率的同时,也令安全性得到了极大的提高;虽然嵌入式隐写平台兼容Windows,但编码效率数倍领先于PC。

This paper proposes a video steganography scheme based on edge features to minimize distortion,and implements it on embedded platform. Intra-prediction steganography mode is selected in this scheme,and the following improvements are made to the algorithm: constructing cost function based on edge features to reduce the damage to spatial correlation in the hard-to-detect edge areas of encrypted data sets;introducing a minimum distortion framework to measure steganographic distortion to solve the problem of insufficient distortion quantization and improve security;citing in the coding and decoding process. The mapping rule based on edge strength is introduced to improve the embedding efficiency and reduce the prediction bias. Finally,experiments show that the steganographic scheme proposed and designed not only guarantees video quality and coding efficiency,but also greatly improves security. On the other hand,although the embedded steganography platform is compatible with Windows,its coding efficiency is several times higher than that of PC.