Demonstration of shift, scale, and rotation invariant target recognition using the hybrid opto-electronic correlator

Gamboa, Julian; Fouda, Mohamed; Shahriar, Selim M.

doi:10.1364/oe.27.016507

Cited by 15 publications

(6 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As compared to the JTC correlator, that also eliminates the recording process of filters, the HOC correlator uses improved materials, that do not require exposure to high voltages that might damage them beyond repair. The HOC correlator also possesses the advantage of being invariant to shift changes and, in more recent developments, architectures of the HOC correlator were developed that are also invariant to scale and rotation changes [ 58 ].…”

Section: Classification and Overview Of The Principles Of Optical Cor...mentioning

confidence: 99%

“…It was proposed as a part of an interferometer prototype, in CMB (cosmic microwave background) applications, for the characterization of the lower frequency bands (approximately between 10 and 12 GHz) of the CMB [ 59 ]. The hybrid opto-electronic correlator was shown to achieve comparable results as the holographic correlator, without the need of using holograms [ 58 ].…”

Section: Classification and Overview Of The Principles Of Optical Cor...mentioning

confidence: 99%

“…Various architectures of hybrid opto-electronic (HOC) correlators have also been proposed. Gamboa et al [ 58 ] obtained target recognition with invariance to scale, shift and rotation using a HOC correlator. First, an opto-electronic combination of signal processing was employed to obtain the PMT (polar Mellin transformed) versions of two images.…”

Section: Recent Applications and Implementations Of Optical Correlatorsmentioning

confidence: 99%

See 2 more Smart Citations

Optical Correlators for Cryptosystems and Image Recognition: A Review

Drăgulinescu

2023

Sensors

View full text Add to dashboard Cite

Optical correlators are efficient optical systems that have gained a wide range of applications both in image recognition and encryption, due to their special properties that benefit from the optoelectronic setup instead of an all-electronic one. This paper presents, to the best of our knowledge, the most extensive review of optical correlators to date. The main types are overviewed, together with their most frequent applications in the newest contributions, ranging from security uses in cryptosystems, to medical and space applications, femtosecond pulse detection and various other image recognition proposals. The paper also includes a comparison between various optical correlators developed recently, highlighting their advantages and weaknesses, to gain a better perspective towards finding the best solutions in any specific domain where these devices might prove highly efficient and useful.

show abstract

Section: Classification and Overview Of The Principles Of Optical Cor...mentioning

confidence: 99%

Section: Classification and Overview Of The Principles Of Optical Cor...mentioning

confidence: 99%

Section: Recent Applications and Implementations Of Optical Correlatorsmentioning

confidence: 99%

See 1 more Smart Citation

Optical Correlators for Cryptosystems and Image Recognition: A Review

Drăgulinescu

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…This system combines a 2D optical image correlator with a 1D atomic temporal correlator, which allows it to perform correlations efficiently in the spatial and temporal domains simultaneously. Optical spatial correlators, such as the hybrid opto-electronic correlator (HOC) [3][4][5][6] , have recently been shown to perform spatial scale, shift, and rotation invariant (SSRI) image correlation at a speed of over 700 frame per second (fps) via the 2D polar Mellin transform (PMT). Similarly, the 1D Mellin transform (MT) can be applied to the time frequency axis to enable frequency-shift invariant temporal correlation, allowing events that occur at different speeds to be recognized.…”

Section: Introductionmentioning

confidence: 99%

Investigation of frequency invariance in automated event recognition using resonant atomic media

Shen,

Gamboa,

Hamidfar

et al. 2024

Practical Holography XXXVIII: Displays, Materials, and Applications

View full text Add to dashboard Cite

Coherent excitation of a resonant medium yields a nonlinear response to the Fourier spectrum of the input signals. This property can be exploited to produce a 1D temporal correlator by applying two signals simultaneously, and subsequently reading out the state of the medium. This intricate process of nonlinear responses generates multiple time-delayed outputs, where we are only interested in the specific segment that pertains to the cross-correlation. To this end, the Schrödinger equation is used as a model to accurately determine the precise time code and location of the desired output. Here, we show via simulations how this may be used for 1D event recognition. By comparing a reference signal to a query signal, we can expect a prominent peak in the cross-correlation if there is a match. Such a system is inherently delay-invariant due to the properties of the Fourier transform but is not invariant to scaling in the time-domain (i.e., frequency shifting). We additionally show how frequency-shift invariant correlation can be achieved by pre-processing the input signals via the Mellin transform. This technique is tested using audio signals to achieve speech recognition, where invariance to frequency shifts means that individual phrases may be recognized independently of the voice of the speaker. This approach can be extended to three-dimensional video recognition systems for real-time event recognition. By utilizing the frequencyshift invariant technique, the system can effectively correlate videos with different time scales, making it applicable to various fields, such as surveillance and copyright plagiarism detection.

show abstract

“…In other research 25 , the authors proposed hybrid opto-electronic correlator (HOC) with spatial light modulators (SLMs), detector arrays, and field programmable gate arrays (FPGAs) for shift invariant target recognition. Due to the speed limitation caused by the serial communication between the devices within the HOC, the authors in 26 proposed a system called Integrated Graphic Processing Unit (IGPU) to increase the processing speed of HOC. However, it is obvious that the usage of application-specific integrated circuits (ASIC) limits the system scalability for big data processing.…”

Section: Introductionmentioning

confidence: 99%

Translation-invariant optical neural network for image classification

Sadeghzadeh

Koohi

2022

Sci Rep

View full text Add to dashboard Cite

The classification performance of all-optical Convolutional Neural Networks (CNNs) is greatly influenced by components’ misalignment and translation of input images in the practical applications. In this paper, we propose a free-space all-optical CNN (named Trans-ONN) which accurately classifies translated images in the horizontal, vertical, or diagonal directions. Trans-ONN takes advantages of an optical motion pooling layer which provides the translation invariance property by implementing different optical masks in the Fourier plane for classifying translated test images. Moreover, to enhance the translation invariance property, global average pooling (GAP) is utilized in the Trans-ONN structure, rather than fully connected layers. The comparative studies confirm that taking advantage of vertical and horizontal masks along GAP operation provide the best translation invariance property, compared to the alternative network models, for classifying horizontally and vertically shifted test images up to 50 pixel shifts of Kaggle Cats and Dogs, CIFAR-10, and MNIST datasets, respectively. Also, adopting the diagonal mask along GAP operation achieves the best classification accuracy for classifying translated test images in the diagonal direction for large number of pixel shifts (i.e. more than 30 pixel shifts). It is worth mentioning that the proposed translation invariant networks are capable of classifying the translated test images not included in the training procedure.

show abstract

Demonstration of shift, scale, and rotation invariant target recognition using the hybrid opto-electronic correlator

Cited by 15 publications

References 21 publications

Optical Correlators for Cryptosystems and Image Recognition: A Review

Optical Correlators for Cryptosystems and Image Recognition: A Review

Investigation of frequency invariance in automated event recognition using resonant atomic media

Translation-invariant optical neural network for image classification

Contact Info

Product

Resources

About