The Image Retrieval Based on Attribute Feature Extraction Coupled Firefly Optimization

In order to solve the problem of semantic gap in image retrieval algorithm, and improve the performance of image retrieval, proposed a Image retrieval based on the optimization of regional segmentation coupled firefly algorithm.Firstly, based on the characteristics of pixel clustering, each image was segmented into different regions and the color and texture features were extracted by using the normalized segmentation model; Then, the Mover's distance Earth (EMD) criterion was introduced to calculate the distance between the query image and the database image, get the feedback from the nearest image and the user's request; Then through the SVM to get the image feedback information for classification and learning, and the introduction of firefly algorithm for iterative optimization, Through continuous iteration, relevant and not related to the continuous image filtering, so as to obtain the desired results.Through in the Corel image database experiments show that: compared with the commonly used image retrieval algorithm, this algorithm has higher precision and recall, ANMRR lower value, the efficiency of the algorithm has been improved, can be more accurate to find users expect image.