中文摘要：

针对隐写技术主要存在对失真的感知度较大及处理的图像普遍压缩程度过大的问题，研发了一种新的隐写术，该技术以KILT为前提，引入一种迭代聚类的算法得到所需解，对数据进行分析，利用KLT对需要进行隐藏的图形进行处理，最终使用最低有效位对需要保密的信息进行隐藏处理。需要获取信息时，可以根据原始像素矩阵利用反向线性变换运算的办法获取源图像。实践证明，这种方法无论是从容量方面，还是信噪方面都要优于常用算法，并且获取的图像失真较小。

英文摘要：

The steganography mainly has two problems of overhigh compression degree for the processed image and high perception degree for distortion. To solve these problems, a new steganography based on KLT （Karhunen-Loeve transform） was developed. An iterative clustering algorithm is introduced into the steganography to obtain the needing solution and analyze the data. The KLT is used to process the image which needs to be hidden. The least significant bit is used for hiding processing for the confidential information. When some information needs to be obtained, the reverse linear transform operation approach is used to acquire the source image according to the original pixel matrix. The practice results prove this method is superior to other commonly-used algorithms in the aspects of capacity and signal-to-noise ratio, and can acquire the image with small distortion.