Combining multiple neural nets for visual feature selection and classification

摘要

We present a system for object recognition in real images employing three different types of neural networks, which accomplish feature extraction and -classification. The main advantages of the method are its portability to different objectdomains without extensive parameter adjustments or changes in the feature extraction, and the low computational effort. This is achieved using a combination of vector quantization, principal component analysis and a network for nonlinear classificationtasks.