How neurons learn to associate 2D-views in invariant object recognition
GM Wallis · 1996
A local learning rule is shown to be able to account for the association of images together on the basis of temporal order rather than spatial configuration, as described in single cell recording results published by Miyashita (1988). Possible reasons for requiring such learning are then given in the context of invariant object recognition This work was supported by a Fellowship from the Max-Plank Gesellschaftschaft This document is available as /pub/mpi-memos/TR-37.ps via anonymous ftp from ftp.mpik-tueb.mpg.de or from the World Wide Web, http://www.mpik-tueb.mpg.de/projects/TechReport/list.html. 1 Introduction The primate temporal lobe has long been implicated in the recognition of objects (Ungerleider & Mishkin, 1982; Goodale & Milner, 1992). Single cell recording in this area has revealed neurons responsive to images of faces (Desimone, 1991; Rolls, 1992) and other, more abstract image features (Tanaka et al., 1991; Miyashita & Chang, 1988). Of particular interest is the ability...