Multi-feature intelligent image retrieval model based on edge detection

Two-stage intelligent image retrieval model is proposed. In the first rank the model detects the edge of original image by the means of new Canny operator and color vector angle respectively, and then it conducts image fusion of the two results that produced by those two detection methods based on db2 wavelet, and finally bit error rate is used to compare the similarity of edge information in order to narrow the retrieval range, which turned out a new alternative image library. In the second rank HSI color invariant model is used to improve the color correlation matrix for multi-feature image retrieval with edge histogram. Experimental results show that our model is accurate and efficient in the user-interested images retrieval with better precision and recall.