3Scan

3Scan build a Knife-Edge Scanning Microscope (KESM) that serially sectioned whole organ tissue while imaging the section with a line scanning camera. The tissue was sliced with an ultra-fine diamond knife edge, allowing for a light source to pass through the blade and illuminate the tissue while it was being sliced. The line scanning camera captured this data in lockstep with the position change of the knife, allowing for whole organ image stacks to be collected with minimal need for image registration.

The challenge was that now there were hundreds of terabytes of large scale images that needed to be analyzed, creating value from the extracted insight in the data.

I was the first data analysis hire to 3Scan, and helped build the foundational algorithms that were performed on the image data. This included luminance and color correction across multiple histopathology stains such as H&E or india ink for vascular tracing.

After image preprocessing, one of my primary tasks was image segmentation of all microvasculature, often in voxel volumes of shape 40000 x 30000 x 3000 (anisotropic) pixels, and create 3D maps for visualization, along with guiding a team that created a graphical network model of the organ vasculature based on the segmentation.

It was a joy working with a fast nimble startup team on a technically challenging biomedical product. Being the first data analysis hire and helping to build the rest of the team taught me about team building and system design.
I learned a lot about how to write engineer code rather than scientist code as well, which set me up for success in all my future roles.