Photorefractive two-beam coupling joint transform correlator: modeling and performance evaluation

Abstract

The photorefractive two-beam coupling joint transform correlator combines two features. The first is embedded semi-adaptive optimality, which weighs the correlation against clutter and noise in the input, and the second is the intrinsic dynamic range compression nonlinearity, which improves several metrics simultaneously without metric trade-off. Although the two beam coupling correlator was invented many years ago, its outstanding performance was recognized on only relatively simple images. There was no study about the performance of this correlator on complicated images and using different figures of merit. In this paper, the study is extended to more complicated images. For the first time, to our knowledge, we demonstrate simultaneous improvement in metrics performance without metric trade-off. The performance was evaluated compared to the classical joint transform correlator. A typical experimental result to validate the simulation results was also shown in this work. The best performing operation parameters were identified to guide the experimental work and for future comparison with other well-known optimal correlation filters.

Optical Correlators for Cryptosystems and Image Recognition: A Review

“…To improve the characteristics of the conventional JTC architecture, the PR-TBCJTC was also introduced. It presents five main advantages as compared to the classical JTC: a reduced number of false alarms, leading to a more efficient detection; an operation in real-time, adaptable; an improvement of the SNR (signal-to-noise ratio) by means of dynamic range compression techniques; an enhancement of the correlation peaks intensity at the output; a simultaneous improvement of all the metrics involved [ 41 ].…”

“…Such alternatives to the classical matched filter are: POF (phase-only filter); correlation filters with encoding of both amplitude and phase; trade-off correlation filters; ternary encoded filters; correlation filters with syntactic discrimination; optimal filters etc. [ 41 ]. The names of seven of the most used filters in VLC correlators are: CMF (classical matched filter), CTMF (complex ternary matched filter), POF, BPOF (binary POF), QPF (quad phase filter), MACE (minimum average correlation energy filter) and the composite filter [ 1 ].…”

“…However, regardless of type, most of the matched filters have the advantage of being relatively easy to implement. However, they have the drawback of introducing, as compared to the JTC, an additional processing step, required for phase information extraction [ 41 ]. The quest for finding the best matched filter has been directed towards solving some problems or improving some characteristics of the VLC correlator, such as: better discrimination; lower rate of false alarms; improved SNR; obtaining the invariance of the correlator to distortions caused by scale, rotation, or other changes [ 41 ].…”

Structured light based optical pattern recognition:Principle and technical progress

“…20世纪90年代, Khoury等人 [15][16][17][18][19][20][21][22][23][24][25][26] [27,28] . 2000年Banerjee等人 [29] 使 用新型的光折变聚合物材料实现了边缘增强的JTC, 将非线性效应引入到光学信息处理, 使得全光的相关 器进入新的发展阶段, 除了光折变效应这种弱光非线 性效应以外还可以通过强光作用的非线性效应实现光 学模式识别.…”

“…对于非线性频率下转换 过程, 即自发参量下转换(Spontaneous Parametric Down-Conversion, SPDC), 本课题组对泵浦光进行空 间光场调控, 同时结合傅里叶空间中的量子关联成像 过程, 构建一个量子版本的Vander Lugt匹配滤波相关 器 [32,33] , 实现了非定域的人脸识别. [20] .…”