An application of back-propagation networks to handwritten zipcode recognition is presented. Minimal preprocessing of the data isrequired, but the architecture of the network is highly constrained andspecifically designed for the task. The input of the network consists ofsize-normalized images of isolated digits. The performance on zip codedigits provided by the US Postal Service is 92% recognition, 1%substitution, and 7% rejects. Structured neural networks can be viewedas statistical methods with structure which bridge the gap betweenpurely statistical and purely structural methods
展开▼