Template matching: matched spatial filters and beyond

作者:

Highlights:

摘要

Template matching by means of cross-correlation is common practice in pattern recognition in spite of its drawbacks. This paper reviews some results on how these shortcomings can be removed. Several techniques (Matched Spatial Filters, Synthetic Discriminant Functions, Principal Components Projections and Reconstruction Residuals) are reviewed and compared on a common task: locating eyes in a database of faces. New variants are also proposed and compared: least squares Discriminant Functions and the combined use of projections on eigenfunctions and the corresponding reconstruction residuals. Finally, approximation networks are introduced in an attempt to improve filter design by the introduction of nonlinearity. © I997 Pattern Recognition Society. Published by Elsevier Science Ltd.

论文关键词:Template matching,Correlation,Neural networks,Learning,HyperBF networks,Principal components

论文评审过程:Received 27 November 1995, Revised 1 July 1996, Accepted 15 July 1996, Available online 7 June 2001.

论文官网地址:https://doi.org/10.1016/S0031-3203(96)00104-5