Three-dimensional pattern recognition with a single two-dimensional synthetic reference function

Abstract

A novel, to our knowledge, method of distortion-invariant three-dimensional ͑3-D͒ pattern recognition is proposed. A single two-dimensional synthetic discriminant function is employed as a reference function in the 3-D correlator. Thus the proposed system is able to identify and locate any true-class object in the 3-D scene. Preliminary simulation and experimental results are presented.

Object recognition using three-dimensional optical quasi-correlation

“…[3][4][5][6][7][8]10,11 Rosen has proposed to extend the correlation dimensions from 2-D to 3-D by introducing 3-D optical Fourier transform 3,4 (FT). By fusion of several projections of the tested scene, a 3-D object function is first Fourier transformed, then filtered by some 3-D reference filter, and finally inversely Fourier transformed into the correlation space.…”

“…Later, the present authors proposed a method to implement 3-D distortion invariance for the 3-D pattern recognition by use of a single 2-D synthetic reference function in the 3-D JTC. 6 As an alternative to the 3-D JTC, we also proposed a hybrid 3-D correlator with general complex filters. 7 Esteve-Taboada et al introduced into the 3-D object recognition 8 the Fourier transform profilometry technique.…”

“…The main criticism of all the various versions of the 3-D correlators developed at Ben-Gurion University [3][4][5][6][7] (in the following we term them all BGU correlators) has been directed against their complexity and their extensive use of digital computation. To overcome these drawbacks, we propose here a new 3-D pattern recognition system, termed 3-D optical quasi-correlator (OQC).…”

Optical quasi-three-dimensional correlation

“…In this method the observed scene must be processed with intensive digital algorithms to reconstruct the 3-D image inside the computer memory before any correlation can be employed. Recently other attempts of 3-D optical pattern recognition were reported [2][3][4][5][6][7][8][9] . Rosen has proposed to extend the correlation dimensions from 2-D to 3-D by introducing 3-D optical Fourier transform (FT) 2,3 .…”

“…His method has been firstly demonstrated with a 3-D joint transform correlator (JTC) 4 . Later, we have proposed a method to implement 3-D distortion invariance for the 3-D pattern recognition by use of a single 2-D synthetic reference function in the 3-D JTC 5 . As an alternative to the 3-D JTC, we have also proposed a hybrid 3-D correlator with general complex filters 6 .…”