Parallel Global Optimal Approach of Feedback for 3D CAD Model Retrieval

A parallel global optimal approach of feedback is proposed for 3D CAD model retrieval in this study. First, a novel unified mathematical model with multi-object for similarity field modification based relevance feedback is brought forward and the simplification is also given. Second, a new algorithm based on particle swarm optimization is proposed to optimize above model. In this algorithm, the particle can fly under the guide of its experience and a leader queue helps to avoid immature convergence. At last, an implementation is given and the results show that better feedback results are obtained for CAD model retrieval with the proposed method.