Image Retrieval Based on Improved Supervised LLE Learning Method

In this paper, identification information is put into the distance measure, using this new distance measure instead of the Euclidean distance to construct k -neighbor, We propose a new improved supervised locally linear embedding method. The IS-LLE method can reduce the vectors dimension with keeping their original topology structure into a lower dimension space, the methods increases the margin of classes in the transformed space. Experiment shows that the proposed IS-LLE method can achieve higher precision rate in CBIR.