Summary: Selective Attention for Handwritten
Digit Recognition
Ethem Alpaydin
Department of Computer Engineering
Bo–gazi¸ci University
Istanbul, TR--80815 Turkey
alpaydin@boun.edu.tr
Abstract
Completely parallel object recognition is NP­complete. Achieving
a recognizer with feasible complexity requires a compromise be­
tween parallel and sequential processing where a system selectively
focuses on parts of a given image, one after another. Successive
fixations are generated to sample the image and these samples are
processed and abstracted to generate a temporal context in which
results are integrated over time. A computational model based on a
partially recurrent feedforward network is proposed and made cred­
ible by testing on the real­world problem of recognition of hand­
written digits with encouraging results.
1 INTRODUCTION
For all­parallel bottom­up recognition, allocating one separate unit for each possible

Source: Alpaydýn, Ethem - Department of Computer Engineering, Bogaziçi University

Collections: Computer Technologies and Information Sciences