Algorithm for Object Detection using Multi-Core Parallel Computation

Fuzzy support vector machine (FSVM) and template matching are both widely used in object detection. To improve the computational efficiency, an algorithm combining the FSVM classifier with template matching is proposed and we parallelize the template matching algorithm on a multicore platform with OpenMP. The samples are firstly classified by the template matching, and then refined by the FSVM classifier. A three-tier hierarchical pyramid is used in fast matching and a second-order polynomial kernel function is used in the FSVM classifier. The experimental results show that the proposed parallel algorithm significantly improves the computation performance compared with traditional serial algorithms.