激光与光电子学进展, 2017, 54 (2): 021004, 网络出版: 2017-02-10   

基于数据降维与对称二值模式的图像Hash算法

Image Hash Algorithm Based on Data Dimension Reduction and Symmetric Binary Pattern
作者单位
1 平顶山教育学院计算机系, 河南 平顶山 467000
2 郑州大学信息工程学院, 河南 郑州 450001
摘要
为了解决当前图像Hash算法难以兼顾较高的感知稳健性与篡改识别率的不足, 提出了基于数据投影降维机制与对称局部二值模式的紧凑图像Hash算法。利用双线性插值来预处理图像, 使Hash具有固定的长度; 引入对数极坐标变换, 将其转变为二次图像; 利用Gabor滤波器平滑二次图像; 基于模糊集理论, 设计对称局部二值模式算子, 获取稳健特征; 定义数据投影降维机制与量化规则, 生成紧凑的中间Hash比特序列; 构造一维组合混沌映射, 建立加密模型, 完成比特序列扩散, 以生成图像Hash; 并引入汉明距离, 估算初始图像与接收端图像的Hash相似度, 联合决策阈值, 完成图像认证。测试数据表明, 与当前图像Hash技术相比, 该算法的Hash更紧凑, 且其感知稳健性与敏感性更高。
Abstract
In order to solve the problem of difficulty of both the high perception robustness and tampering identification rate in the current image Hash algorithm, the compact image Hash algorithm based on data projection dimension reduction mechanism and fuzzy symmetric local binary pattern is proposed. The generated Hash has a fixed length by introducing the bilinear interpolation mechanism to preprocess the image. And the pretreatment image is transformed into the secondary image by the log polar transformation. The secondary image is smoothed by Gabor filter. The fuzzy symmetric local binary pattern operator is designed based on the fuzzy theory. And the compact intermediate Hash sequence is got by defining the data projection dimension reduction mechanism. The image Hash is generated by diffusing the bit Hash based on designing the one-dimensional combined chaotic map. The similarity between the original image and the image of the receiving end is estimated by introduction the Hamming distance and decision threshold to finish the authentication of image. Testing data show that this algorithm has stronger perception robust and sensitivity with tighter Hash length than the current image Hash technologies.

王彦超, 郭静博, 周丽宴. 基于数据降维与对称二值模式的图像Hash算法[J]. 激光与光电子学进展, 2017, 54(2): 021004. Wang Yanchao, Guo Jingbo, Zhou Liyan. Image Hash Algorithm Based on Data Dimension Reduction and Symmetric Binary Pattern[J]. Laser & Optoelectronics Progress, 2017, 54(2): 021004.

本文已被 3 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!