-
Super resolution computational saturated absorption microscopy
Authors:
Gabe Murray,
Jeff Field,
Patrick Stockton,
Ali Pezeshki,
Jeff Squier,
Randy Bartels
Abstract:
Imaging beyond the diffraction limit barrier has attracted wide attention due to the ability to resolve image features that were previously hidden. Of the various super-resolution microscopy techniques available, a particularly simple method called saturated excitation microscopy (SAX) requires only a simple modification of a laser scanning microscope where the illumination beam power is sinusoida…
▽ More
Imaging beyond the diffraction limit barrier has attracted wide attention due to the ability to resolve image features that were previously hidden. Of the various super-resolution microscopy techniques available, a particularly simple method called saturated excitation microscopy (SAX) requires only a simple modification of a laser scanning microscope where the illumination beam power is sinusoidally modulated and driven into saturation. SAX images are extracted from harmonics of the modulation frequency and exhibit improved spatial resolution. Unfortunately, this elegant strategy is hindered by the incursion of shot noise that prevents high resolution imaging in many realistic scenarios. Here, we demonstrate a new technique for super resolution imaging that we call computational saturated absorption (CSA) in which a joint deconvolution is applied to a set of images with diversity in spatial frequency support among the point spread functions used in the image formation with saturated laser scanning fluorescence microscope. CSA microscopy allows access to the high spatial frequency diversity in a set of saturated effective point spread functions, while avoiding image degradation from shot noise.
△ Less
Submitted 26 May, 2023;
originally announced May 2023.
-
Neuro-Symbolic AI: An Emerging Class of AI Workloads and their Characterization
Authors:
Zachary Susskind,
Bryce Arden,
Lizy K. John,
Patrick Stockton,
Eugene B. John
Abstract:
Neuro-symbolic artificial intelligence is a novel area of AI research which seeks to combine traditional rules-based AI approaches with modern deep learning techniques. Neuro-symbolic models have already demonstrated the capability to outperform state-of-the-art deep learning models in domains such as image and video reasoning. They have also been shown to obtain high accuracy with significantly l…
▽ More
Neuro-symbolic artificial intelligence is a novel area of AI research which seeks to combine traditional rules-based AI approaches with modern deep learning techniques. Neuro-symbolic models have already demonstrated the capability to outperform state-of-the-art deep learning models in domains such as image and video reasoning. They have also been shown to obtain high accuracy with significantly less training data than traditional models. Due to the recency of the field's emergence and relative sparsity of published results, the performance characteristics of these models are not well understood. In this paper, we describe and analyze the performance characteristics of three recent neuro-symbolic models. We find that symbolic models have less potential parallelism than traditional neural models due to complex control flow and low-operational-intensity operations, such as scalar multiplication and tensor addition. However, the neural aspect of computation dominates the symbolic part in cases where they are clearly separable. We also find that data movement poses a potential bottleneck, as it does in many ML workloads.
△ Less
Submitted 13 September, 2021;
originally announced September 2021.
-
Single-Pixel Fluorescent Diffraction Tomography
Authors:
Patrick A. Stockton,
Jeffery J. Field,
Jeff Squier,
Ali Pezeshki,
Randy A. Bartels
Abstract:
Optical diffraction tomography is an indispensable tool for studying objects in three-dimensions due to its ability to accurately reconstruct scattering objects. Until now this technique has been limited to coherent light because spatial phase information is required to solve the inverse scattering problem. We introduce a method that extends optical diffraction tomography to imaging spatially inco…
▽ More
Optical diffraction tomography is an indispensable tool for studying objects in three-dimensions due to its ability to accurately reconstruct scattering objects. Until now this technique has been limited to coherent light because spatial phase information is required to solve the inverse scattering problem. We introduce a method that extends optical diffraction tomography to imaging spatially incoherent contrast mechanisms such as fluorescent emission. Our strategy mimics the coherent scattering process with two spatially coherent illumination beams. The interferometric illumination pattern encodes spatial phase in temporal variations of the fluorescent emission, thereby allowing incoherent fluorescent emission to mimic the behavior of coherent illumination. The temporal variations permit recovery of the propagation phase, and thus the spatial distribution of incoherent fluorescent emission can be recovered with an inverse scattering model.
△ Less
Submitted 5 August, 2020;
originally announced August 2020.
-
Compressive Raman imaging with spatial frequency modulated illumination
Authors:
Camille Scotté,
Siddharth Sivankutty,
Patrick Stockton,
Randy A. Bartels,
Hervé Rigneault
Abstract:
We report a line scanning imaging modality of compressive Raman technology with spatial frequency modulated illumination using a single pixel detector. We demonstrate the imaging and classification of three different chemical species at line scan rates of 40 Hz.
We report a line scanning imaging modality of compressive Raman technology with spatial frequency modulated illumination using a single pixel detector. We demonstrate the imaging and classification of three different chemical species at line scan rates of 40 Hz.
△ Less
Submitted 11 April, 2019; v1 submitted 21 January, 2019;
originally announced January 2019.