Skip to main content

Looking at microscope slides with computer vision

We are making progress on the auto-scope project! The mechanical aspects of the microscope are all finished and working*. We’ve written most of the code to control the x, y, and z of the microscope and we’re starting to move onto the whole slide stitching and machine learning aspects of the project.

Check out this (oldish) video of the stage moving around. Look at it go!

Read more…

Auto-Scope - Stage Driver V1

I finished the stage driver prototype for our programmable microscope project: Auto-Scope.

The design is pretty simple; there are two mounting points on the stage that I chose to use to support the unit. The unit is designed to be modular for easy printing and upgrading. I had to print a few parts along the way while I iterated on the design, but it's now looking pretty good. The final design is currently printing on my CR-10 and will hopefully come out as depicted above. Once it's assembled the real fun (aka the Python) can start!

Read more…

Auto-Scope

At the last meeting, the group decided that python-friends should take a more hands-on project based turn. The nominate project is to build an open source whole slide scanner and automated cell discrimination and counting system. This project will involve many aspects of Python programming from controlling motors to building deep learning classifiers. We will work on this project at the weekly meetings. All are welcome to contribute, and no previous background is required.

Follow our progress on this blog.

Counting 80,000+ blood cells by hand will drive anyone to search for alternatives. A recent project I worked on involved looking for parasites in freshwater turtle blood cells (lots of turtle blood cells...). Due to the exotic nature of turtle cells, there are no readily available cell counting systems, and with the prospect of counting another 100,000+ cells I think we need to make one. Luckily we have Python super powers!

The first part of the project involves designing motor mounts to control the stage and focus. I'm going to work on this over the weekend, and hopefully, we'll have a system to start prototyping code on Monday.

Welcome to the blog!

I built this blog using the Python-based static site generator - Nikola. Learn more at the Nikola website.

I used a nifty little Python tool called doctr to connect the blog to Travis CI. Now fully equipped with continuous integration the site will be rebuilt and deployed with each push to the src branch.

Hopefully you'll see the build passing here...