Abstract:

The application of data mining clustering algorithms in contentbased image retrieval can effectively optimize the retrieval speed and effect, to be more specific, fuzzy clustering algorithms fit better the fuzzy characteristics of image retrieval, but affect the retrieval function with a long clustering time, so a color image retrieval method based on improved blocked color histograms and fuzzy cmeans clustering is  proposed. First, each image in the image library is blocked, the improved color characteristic information of each block is extracted; a fuzzy cmeans clustering algorithm is used to cluster color feature vectors, and each cluster center of image class is obtained; finally, the similarity between the sample image and the corresponding categories is calculated, returning the retrieval results according to the size of similarity. The experiments show that the proposed method has a  higher recall rate and a higher precision rate, and less feature dimension of extraction, a shorter clustering time and a quicker retrieval speed.

Key words: color image retrieval;image blocking;improved color histogram;fuzzy cmeans clustering algorithm