Abstract

An image sequence-based framework for appearance-based ob ject recognition is proposed in this paper. Compared with the meth- ods of using a single view for ob ject recognition, inter-frame consistencies can be exploited in a sequence-based method, so that a better recog- nition performance can be achieved. We use the nearest feature line method (NFL) [8] to model each ob ject. The NFL method is extended in this paper by further integrating motion-continuity information between features lines in a probabilistic framework. The associated recognition task is formulated as maximizing an a posteriori probability measure. The recognition problem is then further transformed to a shortest-path searching problem, and a dynamic-programming technique is used to solve it.