Abstract:

Accurate image retrieval is required to index and retrieve large number of images from huge databases. This paper presents an efficient approach to encode the color and textural features of images from the local 2×2 pixel grids. The color features are extracted by quantizing the HSV color space into a single channel with reduced number of shades. The texture information is encoded with 5 structuring patterns generated from the locally structured elements chosen as a basis. Color and textural features are fused together to construct the inherent color and texture hybrid image descriptor (CTHD) which is rotation and scale invariant. This fusion is carried out by extracting textural cues over each shade independently. CTHD has been tested on the Corel-1K and Corel-5K datasets, and the experimental results suggest that CTHD outperforms state-of-the-art descriptors. The performance of the CTHD can also be effectively used under more complex image transformations.

Key words: color feature, textural feature, structure element descriptor, rotation and scale invariant

摘要：

如何从数据庞大的图像数据库中建立索引并准确检索出相关图像是现阶段急需解决的问题。从图像的局部2×2像素网格提取图像的颜色和纹理特征，其中颜色特征从经过非均匀量化的HSV颜色空间提取，定义了5类新颖的纹理结构描述符用来描述图像的纹理。通过分别提取每个颜色值的纹理特征结构得到最终的具有旋转和尺寸不变性的颜色和纹理混合特征描述符（color and texture hybrid descriptor，CTHD）。在Corel标准图片库和Corel-5K图片库中进行了实验，经过与其他图像检索方法对比，所提方法具有更好的检索准确率。

关键词: 颜色特征, 纹理特征, 结构元素描述符, 旋转和尺寸不变性