Aspect graph enables new and exciting applications, but also poses significant technical challenges. In this paper, we give a brief overview of aspect graph and its applications with an emphasis on 3D object recognition at first. We then compare various partition algorithms for aspect graph space and analyze their advantages and disadvantages; furthermore we give two improving methods to aspect graph space partition: pruning strategy and approximating approach. Finally, based on a specific application on airplane recognition, we propose a 3D object recognition strategy using multi-resolution aspect graph, and the result shows that it can effectively solve the problems of practicability and stability in actual 3D object recognition.
展开▼