A5058629781- Digital Holography and Microscopy
- Cell Image Analysis Techniques
- Image Processing Techniques and Applications
- Advanced Fluorescence Microscopy Techniques
- Neural Networks and Reservoir Computing
- Photonic and Optical Devices
- Sparse and Compressive Sensing Techniques
- Optical Network Technologies
- AI in cancer detection
- Random lasers and scattering media
- Optical Coherence Tomography Applications
- Advanced Optical Imaging Technologies
- Advanced X-ray Imaging Techniques
- Photoacoustic and Ultrasonic Imaging
- Optical measurement and interference techniques
- Molecular Biology Techniques and Applications
- Microwave Imaging and Scattering Analysis
- Advanced Image Processing Techniques
- Advanced Electron Microscopy Techniques and Applications
- Digital Imaging for Blood Diseases
- Advanced Vision and Imaging
- Blind Source Separation Techniques
- Microfluidic and Bio-sensing Technologies
- Optical Polarization and Ellipsometry
- Terahertz technology and applications
University of California, Los Angeles
2016-2025
California NanoSystems Institute
2017-2025
Samueli Institute
2021-2023
Bioengineering Center
2019-2022
Roche (United States)
2021
Duke Medical Center
2021
United States Food and Drug Administration
2021
Office of Science
2021
Rutgers, The State University of New Jersey
2021
Taipei Veterans General Hospital
2021
We introduce an all-optical Diffractive Deep Neural Network (D2NN) architecture that can learn to implement various functions after deep learning-based design of passive diffractive layers work collectively. experimentally demonstrated the success this framework by creating 3D-printed D2NNs learned handwritten digit classification and function imaging lens at terahertz spectrum. With existing plethora 3D-printing other lithographic fabrication methods as well spatial-light-modulators,...
Phase recovery from intensity-only measurements forms the heart of coherent imaging techniques and holography. In this study, we demonstrate that a neural network can learn to perform phase holographic image reconstruction after appropriate training. This deep learning-based approach provides an entirely new framework conduct by rapidly eliminating twin-image self-interference-related spatial artifacts. network-based method is fast compute reconstructs amplitude images objects using only one...
Information security and authentication are important challenges facing society. Recent attacks by hackers on the databases of large commercial financial companies have demonstrated that more research development advanced approaches necessary to deny unauthorized access critical data. Free space optical technology has been investigated many researchers in information security, encryption, authentication. The main motivation for using optics photonics is waveforms possess complex degrees...
We demonstrate that a deep neural network can significantly improve optical microscopy, enhancing its spatial resolution over large field-of-view and depth-of-field. After training, the only input to this is an image acquired using regular microscope, without any changes design. blindly tested learning approach various tissue samples are imaged with low-resolution wide-field systems, where rapidly outputs remarkably better resolution, matching performance of higher numerical aperture lenses,...
Abstract Using a deep neural network, we demonstrate digital staining technique, which term PhaseStain, to transform the quantitative phase images (QPI) of label-free tissue sections into that are equivalent brightfield microscopy same samples histologically stained. Through pairs image data (QPI and corresponding images, acquired after staining), train generative adversarial network effectiveness this virtual-staining approach using human skin, kidney, liver tissue, matching stained with...
Holography encodes the three dimensional (3D) information of a sample in form an intensity-only recording. However, to decode original image from its hologram(s), auto-focusing and phase-recovery are needed, which general cumbersome time-consuming digitally perform. Here we demonstrate convolutional neural network (CNN) based approach that simultaneously performs significantly extend depth-of-field (DOF) holographic reconstruction. For this, CNN is trained by using pairs randomly de-focused...
Abstract Recent advances in deep learning have given rise to a new paradigm of holographic image reconstruction and phase recovery techniques with real-time performance. Through data-driven approaches, these emerging overcome some the challenges associated existing methods while also minimizing hardware requirements holography. These recent open up myriad opportunities for use coherent imaging systems biomedical engineering research related applications.
Deep learning has been transformative in many fields, motivating the emergence of various optical computing architectures. Diffractive network is a recently introduced framework that merges wave optics with deep-learning methods to design neural networks. Diffraction-based all-optical object recognition systems, designed through this and fabricated by 3D printing, have reported recognize hand-written digits fashion products, demonstrating inference generalization sub-classes data. These...
Optical machine learning offers advantages in terms of power efficiency, scalability and computation speed. Recently, an optical method based on Diffractive Deep Neural Networks (D2NNs) has been introduced to execute a function as the input light diffracts through passive layers, designed by deep using computer. Here we introduce improvements D2NNs changing training loss reducing impact vanishing gradients error back-propagation step. Using five phase-only diffractive numerically achieved...
Pathology is practiced by visual inspection of histochemically stained slides. Most commonly, the hematoxylin and eosin (H&E) stain used in diagnostic workflow it gold standard for cancer diagnosis. However, many cases, especially non-neoplastic diseases, additional "special stains" are to provide different levels contrast color tissue components allow pathologists get a clearer picture. In this study, we demonstrate utility supervised learning-based computational transformation from H&E...
Mobile-phones have facilitated the creation of field-portable, cost-effective imaging and sensing technologies that approach laboratory-grade instrument performance. However, optical interfaces mobile-phones are not designed for microscopy produce spatial spectral distortions in microscopic specimens. Here, we report on use deep learning to correct such introduced by mobile-phone-based microscopes, facilitating production high-resolution, denoised colour-corrected images, matching...
We report a deep learning-enabled field-portable and cost-effective imaging flow cytometer that automatically captures phase-contrast color images of the contents continuously flowing water sample at throughput 100 mL/h. The device is based on partially coherent lens-free holographic microscopy acquires diffraction patterns micro-objects inside microfluidic channel. These are reconstructed in real time using learning-based phase-recovery image-reconstruction method to produce image each...
Diffractive networks encode the spatial information of objects into power spectrum to classify images with a single-pixel detector.
Abstract Recent advances in deep learning have been providing non-intuitive solutions to various inverse problems optics. At the intersection of machine and optics, diffractive networks merge wave-optics with design task-specific elements all-optically perform tasks such as object classification vision. Here, we present a network, which is used shape an arbitrary broadband pulse into desired optical waveform, forming compact passive engineering system. We demonstrate synthesis different...
Abstract Imaging through diffusers presents a challenging problem with various digital image reconstruction solutions demonstrated to date using computers. Here, we present computer-free, all-optical method see random at the speed of light. Using deep learning, set transmissive diffractive surfaces are trained all-optically reconstruct images arbitrary objects that completely covered by unknown, phase diffusers. After training stage, which is one-time effort, resulting fabricated and form...
A plethora of research advances have emerged in the fields optics and photonics that benefit from harnessing power machine learning. Specifically, there has been a revival interest optical computing hardware, due to its potential advantages for learning tasks terms parallelization, efficiency computation speed. Diffractive Deep Neural Networks (D2NNs) form such an framework, which benefits deep learning-based design successive diffractive layers all-optically process information as input...
An efficient method and system for compressive sensing of hyperspectral data is presented. Compression efficiency achieved by randomly encoding both the spatial spectral domains datacube. Separable architecture used to reduce computational complexity associated with a large volume data, which typical imaging. The enables optimizing ratio between compression ratios. demonstrated simulations performed on real data.
Compressive sensing is a relatively new measurement paradigm which seeks to capture the "essential" aspects of high-dimensional object using as few measurements possible. In this work we demonstrate successful application compressive framework digital Fresnel holography. It shown that when applying approach fields special sampling scheme should be adopted for improved results.
Compressive imaging (CI) is a natural branch of compressed sensing (CS). Although number CI implementations have started to appear, the design efficient system still remains challenging problem. One main difficulties in implementing that it involves huge amounts data, which has far-reaching implications for complexity optical design, calibration, data storage and computational burden. In this paper, we solve these problems by using two-dimensional separable operator. By so doing, reduce...
Optical computing provides unique opportunities in terms of parallelization, scalability, power efficiency, and computational speed has attracted major interest for machine learning. Diffractive deep neural networks have been introduced earlier as an optical learning framework that uses task-specific diffractive surfaces designed by to all-optically perform inference, achieving promising performance object classification imaging. We demonstrate systematic improvements networks, based on a...
We demonstrate the effectiveness of nonlocal means (NLM) filter for speckle denoising in digital holography. The noise adapted version NLM is compared with other common filters and found to give better visual quantitative results.