Real-time Tracking Performance Research Articles

Augmented reality application based on hand shape interaction and palmprint recognition

This paper proposed and implemented an Augmented Reality (AR) system based on hand shape interaction and palmprint recognition. The paper proposed a novel fast identification method of hand contour for contour matching and its movement. This system used an improved ant colony clustering algorithm to get hand center quickly,established relevant rendering system,and registered the virtual object. Meanwhile,for personalized applications,the paper proposed fast Harris palmprint identification algorithm applied in augmented reality. Experiments show that the new algorithm has good recognition precision and real-time tracking performance,and can adapt to the application requirements of augmenting reality systems.

Method for underwater target tracking based on an interacting multiple model

According to the requirements of real-time performance and reliability in underwater maneuvering target tracking as well as clarifying motion features of the underwater target, an interacting multiple model algorithm based on fuzzy logic inference (FIMM) is proposed. Maneuvering patterns of the target are represented by model sets, including the constant velocity model (CA), the Singer model, and the nearly constant speed horizontal-turn model (HT) in FIMM technology. The simulation results show that compared to conventional IMM, the reliability and real-time performance of underwater target tracking can be improved by FIMM algorithm.