基于GM(1,1)模型的压缩域图像表示与检索技术

在讨论基于灰色模型GM(1，1)和Hilbert空间填充曲线的图像压缩算法的基础上，提出了一种新的彩色图像表示与检索技术。它使用GM(1，1)对图像的像素值进行模型化处理，并且在变换域中对这些模型参数进行分析与处理，据此抽取出一种新的图像特征。这种压缩域特征描述了图像的局部细节变化，这种特点正是全局检索方法(例如颜色直方图方法)所小具备的。实验结果表明，这种压缩域图像检索技术能获得较满意的检索质量与性能，其检索结果能较好地同人们的视觉感知结果保持一致。

GM(1,1)-based Approach to Image Representation and Retrieval

Based on the research of image compression using GM(1,1) and Hilbert space-filling curve, a novel approach to content-based image retrieval is proposed. The proposed approach is based on the grey model GM(1,1) and Hilbert space filling curve to model the pixels in an image, and a new image feature is easily extracted by analyzing the model parameters { }in the transformed domain. As the feature captures information about the local variation of intensity values or colors in an image, it is an effective method to tackle the main drawback of the global methods (e.g., color histogram). Experimental results are provided indicating that the proposed approach performs well and gives very good retrieval efficiency, and the quality of image retrieval is in good accordance with human perception.