Imaging and Visualization Workshop 2020 Trainee Spotlight


October 16, 2020
Virtual Event

Featured Trainees

This is an accordion element with a series of buttons that open and close related content panels.

"The effect of electron microscopy image characteristics on the performance of machine learning models" by Jingrui Wei

Modern electron microscopes can resolve single atoms in images of materials, and several convolutional neural networks (CNNs) have been developed to identify atom positions quickly in those images even corrupted by noise, distortion, and poor contrast. In order to bring the models to a more generalized and reusable stage, we have developed a benchmark data set and a set of model performance metrics for this problem covering a realistic range of image quality and used them to evaluate two existing CNN models. We find different performance for metrics including the pixel size, contrast, and image resolution, pointing to the need for community agreement on benchmark tests for these and related imaging tools.

"Studying the early embryonic heartbeat with cardiac light sheet microscopy" by Anjalie Schlaeppi

In Jan Huisken’s lab, we develop and build light sheet microscopes to study development. One of our focus is how the zebrafish embryonic heart pumps and whether we can use it as a source of inspiration for bioengineering projects.

"Using scattered light to improve cellular imaging of the retina" by Ben Sajdak

A custom microscope was developed that scans intact tissue and collects scattered light distribution at every image pixel, allowing digital masks to be applied after image collection. Known and novel methods of detecting intrinsic cellular contrast with this system can be directly applied to improve imaging modalities used in adaptive optics ophthalmoscopy.

"Autofluorescence imaging is sensitive to metabolic changes due to cancer treatments" by Amani Gillette

Label-free autofluorescence imaging of metabolic co-enzymes NADH and FAD is a powerful tool to assess the effects of cancer treatments on cellular metabolism. Here we show that autofluorescence measurements taken after cells are treated with cancer therapies correlate with changes in gold-standard measures of cellular metabolism.

"What can microscopy tell us about fatal pulmonary fibrosis?" by Darian S. James

Idiopathic pulmonary fibrosis (IPF) is a fatal lung disease with unknown pathological etiology and while increased collagen is the defining feature of fibrosis, the underlying remodeling of collagen in the extracellular matrix remains understudied. To this end, we used second harmonic generation (SHG) microscopy and SHG combined with circular dichroism (SHG-CD) to probe the collagen architecture of the fibrotic scar tissue, revealing large structural differences between normal and IPF human lung tissues.

"Collagen characterization in clinical histopathology using convolutional neural networks" by Bin Li

We use convolutional neural networks to directly extract information about collagen fiber arrangement and alignment on clinically standard H&E bright-field images by mapping H&E images to SHG images. Collagen fiber analysis algorithms designed for SHG images can be applied when only H&E images are available.

"Optical Imaging of Collagen Fiber Damage to Assess Thermally Injured Human Skin" by Alex Schroeder

The severity of thermal skin injuries is assessed visually by a clinician early after injury but is accurate only up to 70% of the time among experienced surgeons and little is known about the role of collagen due to lack of technology. We will show a fluorophore?tagged collagen?mimetic peptide that can detect damaged collagen.

"Experimental Visualization & Quantification of Left Ventricular Hemodynamics for Validation of Multi-Venc 4D Flow MRI Using Tomographic Particle Image Velocimetry" by James Rice

Four-dimensional (4D) flow MRI with multi-velocity (Venc) encoding aims to improve velocity-to-noise ratio and the dynamic velocity range of 4D flow MRI, but largely remains unvalidated. The aim of this study was to develop an MRI- and PIV-compatible left ventricle model and tomographic-PIV (tomo-PIV) protocol to acquire a reference dataset to compare against 4D flow MRI measurements obtained with multi-Venc settings. Preliminary flow data were acquired and velocity and vorticity were quantified, demonstrating the effectiveness of the tomo-PIV setup as a first step towards enhanced validation of multi-Venc 4D flow MRI measurements in the LV.

"Ehancing human vision beyond the visible spectrum" by Bryan Rubio Perez

We propose the design of an optical device that passively converts ultraviolet light rays to visible light rays. We use perovskite-nanocrystals suspended in a rubber film to frequency convert the light and implement the material system into a telescope. Our passive device does not require bulky electrical or optical components, or a UV sensitive detector.

"Augmented Lagrangian Digital Image/Volume Correlation" by Jin Yang

Digital image/volume correlation (DIC/DVC) is a powerful, non-invasive experimental method for extracting 2D and 3D volumetric full-field deformation information. The basic idea of this method is to compare images of an object painted with a speckle pattern before and after deformation, and thereby to compute displacements and strains. Here we present a new hybrid algorithm, the augmented Lagrangian digital image/volume correlation (AL-DIC/DVC) which has higher accuracy and behaves more robustly compared to current other local and finite element based global methods.


For Trainee Presenters

Trainees and students: RSVP to present your research lightning talk.  Space is limited and will be confirmed on a first-come, first-served basis.

The lightning talk session is open to UW-Madison postdocs, graduate, and undergraduate students.

Instructions, templates and format will be shared upon receipt of your RSVP. Lightning talks are two-minute, pre-recorded videos with one to two slides.

Presentations are due on Monday, October 12, 2020 at 9am.

Trainee Spotlight RSVP