一种基于HSV颜色空间模糊量化的图像检索方法

Image Retrieval Method Based on Color Fuzzy Quantization in Hue-Saturation-Value

目的:为了提高图像检索的准确率,选择一个合适的颜色空间和准确的量化方法。方法:选取HSV颜色空间,在颜色量化时引入了模糊量化方法,并在此基础上采用分块直方图描述方法进行了相应的实验。结果:实验结果表明,HSV颜色空间模糊量化方法能较好地描述图像的颜色特征,检索结果比较令人满意。结论:该方法能够提高图像检索的准确率,且检索出的相关图像排序更靠前。

Objective To improve accuracy of the image retrieval and choose an appropriate color space and an accurate quantization method based on color feature,which is widely used in content-based image retrieval.Methods An improved color fuzzy quantization method in Hue-Saturation-Value（HSV） color space was proposed.On this basis,block color histogram was extracted as a retrieval feature.Results The color features of images could be efficiently expressed by the color fuzzy quantization in HSV,and its retrieval results were satisfied.Conclusion The accuracy of the image retrieval can be improved and the rank of the relevant image is former.