2018
DOI: 10.1109/jphot.2018.2858568
|View full text |Cite
|
Sign up to set email alerts
|

A Novel Method to Realize Optical Correlation Recognition Based on Neural Network

Abstract: A neural network model is constructed according to 4F optical correlator. New optical filter is obtained by training the neural network. For recognition of distortion targets, the correlation peaks from correlator applying new filter are averagely increased up to 306.7% than applying the OTSDF filter, and correlation peaks are higher, sharper and easier to be distinguished, which indicates optical correlator with new filter have better performance in distortion invariant recognition.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

1
5
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
5

Relationship

1
4

Authors

Journals

citations
Cited by 6 publications
(6 citation statements)
references
References 24 publications
1
5
0
Order By: Relevance
“…The period of the fabricated terahertz wideband filter was 127.19 (53.00 + 74. 19) μm, which is close to the design value of 127.90 μm and is within the error tolerance. The duty ratio of the fabricated terahertz wideband filter was 0.417 (53.00/127.…”
Section: Production and Performance Testingsupporting
confidence: 78%
See 2 more Smart Citations
“…The period of the fabricated terahertz wideband filter was 127.19 (53.00 + 74. 19) μm, which is close to the design value of 127.90 μm and is within the error tolerance. The duty ratio of the fabricated terahertz wideband filter was 0.417 (53.00/127.…”
Section: Production and Performance Testingsupporting
confidence: 78%
“…The duty ratio of the fabricated terahertz wideband filter was 0.417 (53.00/127. 19) and the design value is 0.414, which is within the error tolerance.…”
Section: Production and Performance Testingmentioning
confidence: 68%
See 1 more Smart Citation
“…As compared to a traditional OTSDF (optimized trade-off synthetic discriminant function) filter, an up to 1402.4% increase of the correlation peak index was observed. As compared to the NNCRF (neural network correlation recognition filter), proposed more recently by [ 113 ] (which in turn had obtained a 306.7% increase as compared to the OTSDF filter), the NNRRF achieved better performance in terms of the recognition ability invariant to distortions. The proposed filter was used in a planar integrated 2f optical correlator and was validated by experiments in a real setup.…”
Section: Recent Applications and Implementations Of Optical Correlatorsmentioning
confidence: 97%
“…It enables the optical correlator to easily recognize the distortion target, and each index is greatly improved, but in terms of rotational distortion, it can only recognize the left and right rotations distortion of 50 degrees according to the existing reports. [16] In conclusion, optical correlators applying existing filters still have difficulty in recognizing large rotational angle distortion targets An optical filter is a phase-type device consisting of multiple phase points. The key to the design is to solve the optimal combination of phase points.…”
Section: Introductionmentioning
confidence: 99%