A method for evaluating partially destroyed object´s recognition using an appearance-based model

In this paper, we introduce a new recognition strategy in an eigenspace method that can be used for the recognition of an object´s pose instead of using minimum length principle (MLP). This strategy can also evaluate an object or partially destroyed object whether it can be recognized within this manifold or not. This study also proposes an average-appearance eigenspace method where some sort of destroyed/occluded object´s pose can be represented and recognized.