Senior Design Projects

ECS193 A/B Winter & Spring 2022

Image-centric software development supporting novel slide-free microscopy

Email **********
Richard Levenson, MD
UC Davis Health, Dept. of Pathology and Lab. Med.

Project's details

Image-centric software development supporting novel slide-free microscopy
Our laboratory (located at the Sacramento Med. Ctr.) has developed novel microscopy techniques that can acquire diagnostic-quality images directly from the face of biopsy tissues—without requiring the complex and time-consuming process of standard histology, typically occurring overnight. The latter involves tissue fixation, dehydration, paraffin-embedding, preparation of micron-thick slices, transfer to glass slides, staining, coverslipping and delivery to the diagnostic pathologist. And then these slides still require digital scanning to be converted into digital form suitable for sharing and interpretation via artificial intelligence. Instead, we have developed methods that bypass that and deliver intrinsically digital images within 3 minutes of receiving a piece of tissue. This can greatly accelerate diagnoses, decrease costs and permit rapid reflex testing, knocking days off the patient care experience. A start-up, Histolix, Inc. has been spun off from IP developed at UCD.
Tasks: While we have developed rapid scanning techniques, management of the resulting (large) images require considerable improvement. The tasks include: • improving the user interface • managing image display • providing control over downstream under-the-hood image processing (for example, smart sharpening, local contrast enhancement, recoloring for optimal appearance) • user-control over display parameters (brightness, contrast, color) • AI-based re-coloring—in collaboration with a BMEGG graduate student working on this • Real-time implementation on displayed regions as well as batch-scale processing of entire images (gigabytes) in the background • Down the road, we hope to have molecular image data acquired with the same technology. The information needs to be displayed on top of the standard histology images.
We have a prototype viewer that allows for display and image processing—this needs to be recast as a stand-alone platform. Research components involve exploring a variety of image processing approaches and testing whether a promising Python-based open-source image display and analysis package (napari—developed in the Chan-Zuckerberg center) can serve as a good basis for scientific image exploration.
Useful skillset: • The ideal contributors will have experience with Bitmap structure, parallel computing. • C#,VB, python, .NET. • Knowledge of image processing tools, Open CV, GPU programming is a plus. • Basic data analysis, regression, and optimization tools.
**********
30-60 min weekly or more
Client wishes to keep IP of the project
Attachment Click here
No
Team members N/A
N/A
N/A