基于代表色的图像检索方法研究

Content Based Image Retrieval(CBIR) System Based on Representative Color

文章提出了一种基于颜色的图像检索方法。该方法采用HSI颜色模型对颜色空间进行量化提取代表色,通过对图像进行分块,统计各方块的代表色直方圆,并对各方块代表色直方图赋子不同的权值进行相似性计算,最后给出了相应的相似度计算方法。实验证明该方法能有效地降低检索时间、提高检索效果。

This paper gives an image retrieval method. It uses HSI color model to acquire representative color by quantizing color model, and divides image to get representative color histogram in each region. We can search similar images from the image repository by giving different parameters of color histogram in each segment. The results of experiments show the method can reduce retrieval time and improve retrieval precision.