A new distortion-invariant optical correlation filter to produce easily detectable correlation peaks in the presence of noise and clutter and to provide better intra-class recognition is presented. The new Minimum Noi...
详细信息
作者：
Casasent, DavidCarnegie Mellon University
Center for Excellence in Optical Data Processing Department of Electrical and Computer Engineering PittsburghPA15213 United States
Correlation filters with sharp delta-function correlation peaks [such as phase-only filters and minimum average correlation energy (MACE) filters] do not recognize images on which they are not trained. We show that th...
详细信息
Correlation filters with sharp delta-function correlation peaks [such as phase-only filters and minimum average correlation energy (MACE) filters] do not recognize images on which they are not trained. We show that the MACE filter cannot always recognize intermediate images of true class objects (e.g., aspect views or rotations midway between two training images). New Gaussian-MACE filters offer a solution to this problem.
New distortion-invariant correlation filters for in-plane rotation invariance are considered. These use circular-harmonic functions combined with minimum-average correlation-plane filter techniques. Various circular-h...
详细信息
New distortion-invariant correlation filters for in-plane rotation invariance are considered. These use circular-harmonic functions combined with minimum-average correlation-plane filter techniques. Various circular-harmonic function shortcomings are quantified.
A new matrix inversion algorithm is described. It provides a meaningful estimate of the inverse A-1 of a matrix A on an analog optical processor in a reduced calculation time (compared to other methods). The new neste...
详细信息
A new matrix inversion algorithm is described. It provides a meaningful estimate of the inverse A-1 of a matrix A on an analog optical processor in a reduced calculation time (compared to other methods). The new nested iterative algorithm has no convergence conditions on the matrix and requires fewer operations than prior iterative neural net and other algorithms.
A synthesis algorithm is presented for generating a neural associative processor with piecewise-hyperspherical decision boundaries. Two important characteristics of the algorithm are that it represents each class with...
详细信息
A synthesis algorithm is presented for generating a neural associative processor with piecewise-hyperspherical decision boundaries. Two important characteristics of the algorithm are that it represents each class with a near-minimum number of hyperspheres and that it has proven convergence properties. Classification results are presented for a three-class 3D distortion-invariant aircraft case study (invariant to changes in position, scale, and in-plane and out-of-plane rotation). The processor gives 98% accuracy.< >
A Ho-Kashyap (H-K) associative processor (AP) is shown to have a larger storage capacity than the pseudoinverse and correlation APs and to accurately store linearly dependent key vectors. Prior APs have not demonstrat...
详细信息
A Ho-Kashyap (H-K) associative processor (AP) is shown to have a larger storage capacity than the pseudoinverse and correlation APs and to accurately store linearly dependent key vectors. Prior APs have not demonstrated good performance on linearly dependent key vectors. The AP is attractive for optical implementation. A new robust H-K AP is proposed to improve noise performance. These results are demonstrated both theoretically and by Monte Carlo simulation. The H-K AP is also shown to outperform the pseudoinverse AP in an aircraft recognition case study. A technique is developed to indicate the least reliable output vector elements and a new AP error correcting synthesis technique is advanced.
Several performance criteria are described to enable a fair comparison among the various correlation filter designs: signal-to-noise ratio, peak sharpness, peak location, light efficiency, discriminability, and distor...
详细信息
Several performance criteria are described to enable a fair comparison among the various correlation filter designs: signal-to-noise ratio, peak sharpness, peak location, light efficiency, discriminability, and distortion invariance. The trade-offs resulting between some of these criteria are illustrated with the help of a new family of filters called fractional power filters (FPFs). The classical matched filter, phase-only filter (POF), and inverse filter are special cases of FPFs. Using examples, we show that the POF appears to provide a good compromise between noise tolerance and peak sharpness.
optical correlators utilizing spatially incoherent light are examined and compared with coherent correlators. Frequency and image domain optical correlator architectures are discussed. The effect of speckle noise and ...
详细信息
A new bidirectional optical associative processor is described for searching a hierarchical database that is stored as an adjacency matrix. The paper discusses how the processor can answer relatively complex queries o...
详细信息
暂无评论