CoLDImage: Contrast and luminance distribution for content-based image retrieval
Qingkun Su, Yan Tu Huang, Jingliang Peng · 2011
With the increasing volumes of digital image data and the rapid development of internet technologies, it becomes vital to efficiently and accurately retrieve inquired images from the vast available data resources. In this context, content-based image retrieval has been intensively researched in the past decades. In this work, we propose to use contrast and luminance distribution, abbreviated as CoLD, to describe the textural and luminance characteristics of a digital image. The CoLD descriptor is rotation invariant and scale invariant. In addition, it is very simple to compute. Experimental results demonstrate that, when used for content-based image retrieval, the CoLD descriptor yields significantly higher retrieval precision when compared with the classical methods including gray level co-occurrence matrix and Hu's seven moment invariants.