A fast object recognition system using object appearances

The paper presents a novel appearance-model based fast object recognition system. The speed is achieved by avoiding much of the image processing required for feature extraction and by compactly representing object models in a subspace. Raw images of an object are transformed into an object model in an eigenvector space after a simple process that compresses the data size using a low pass spatial filter along with an interpolation method. The validity of the proposed scheme is demonstrated by a working model-airplane recognition system.