Rotation-invariant pattern recognition using a vector reference

Publisher Website
Google Scholar
Abstract
In this paper, we extend the method of expansion into circular harmonic functions (CHF) by using vector rather than scalar target signatures. The reference vector consists of correlation products between the reference image and individual harmonics of the CHF expansion of the reference image. The use of vectors furnishes certain advantages including the ability to detect targets in poor signal-to-noise or low-contrast environments.

This publication has 3 references indexed in Scilit: