Mezoscope: a smart, affordable, automated microscope.
Mezoscope: a smart, affordable, automated microscope.
We have developed a new open microscopy software/hardware system; for DIY enthusiasts, students, and high spec for research/academia.
We have developed a new open microscopy software/hardware system; for DIY enthusiasts, students, and high spec for research/academia. Read more
About this project
An Affordable Fully Automated 'Open' Microscope.
We have developed a new affordable automated microscopy system, suitable for both lab/research use, K12 and University education, and for the hobbyist/artist/enthusiast!
The original (V.1) Mezoscope prototype.
The Big Problem...
We needed a microscope that could capture large scale images of samples much larger than current microscopes are designed for, so we needed both automation and flexibility for imaging. However, the current products from Nikon/Olympus/Leica (etc.) were far too expensive (in the $45-100k range), and didn't provide flexibility in software/imaging programming. So, we decided to take advantage of newly available stepper motor controllers (e.g. from Sparkfun, Polulu, etc.) and build our own automated system from scratch. Now, after 3-4 years of work, we have it!
What is 'Open' Microscopy?
Current microscopes are shipped with closed/proprietary software. It works well in general, but also limits creativity and innovation. Our goal is to provide an 'open' and really affordable hardware/software platform for novel (automated) microscopy, and to build a community and resource for new algorithms and programs. You can re-program our software to perform automated tasks, like 3-D image stacking, mosaic tiling, and more advanced methods for automatically identifying unique features in a slide or any 'specimen'.
Ultimately we want to develop a comprehensive website with a repository for contributed code (both from us and the community) and a forum for researchers and enthusiasts. With this open system there is no limit.
Polarized light image of breast tissue specimen. Original image is appx. 20,000 x 12,000 pixels! Taken with our new V.2 prototype. Visit our galleries at Tensortek.com for a gigapixel version of this and many more images.
We started with a personal challenge: how to obtain large-scale microscopy images on my small research budget. I am a bioengineer at a local university, and we had a big research project idea, but could not afford the expensive (>$50k) systems from the major manufacturers, Our whole research plan depended on this capability... So, we decided to 'go for broke' and build our own. Along the way, many excited responses from our colleagues (e.g. Wow! can you build me one?) motivated us to start our company (Tensortek.com) and build what we think is the first ever Open Microscope System. We have used a lot of open software for this design, and feel the need to share.
An Open Microscopy Resource
We believe that concepts in microscopy can improve dramatically by a combined effort -- because ideas get better when they are shared. A main goal of this project is to develop and maintain a shared code Repository (web based), and a web-based Forum for automated microscopy code and imaging/math/algorithms. Right now, as far as we know, this doesn't exist. The closest thing is probably the excellent government sponsored ImageJ software, but there's nothing there for microscope control or automation... It's only about image post-processing.
We want to share our Mezoscope, build it with other enthusiasts and not limit this exciting technology to big research labs. The exciting thing is that now we can make it affordable (fulfill your dream of owning a smart imaging machine that will allow you to tap into your inner scientist or artistic genius).
An image of hip capsule tissue, taken with our new V.2 prototype (orig. image is appx. 8500x5900 pixels...
Image of a region of butterfly wing... I like to think of it as art :)
Some Details: The Basic Design
The Mezoscope consists of a lab-quality microscope, a camera (we are using Pixelink right now, but you can use any camera... a great choice is a Canon DSLR, or even one of Canon's point-and-shoots... heck, you could use an HD webcam if you really want to cut costs...), three stepper motor drivers (right now we are using Polulu), and three NEMA stepper motors. The motors are mounted using custom brackets to the thumbwheels of the X-Y stage and Z-focus knob. The microstepping drivers provide up to 1600 steps/rev, so we can achieve really fine micron-level positioning. We have written a simple but powerful GUI (mostly in VB dotNet right now, but we really want to port this all to Python, if we get funded. this is a priority). The current GUI has buttons allowing user control, image preview, and some functional tasks like mosaic imaging (grabbing a grid of images), Z-stacking (grabbing 3-D images of 'thick' specimens), and autofocusing. We are also working on a Python plugin-architecture to allow you to program your own scripts and imaging algorithms.
Basic schematic of the Mezoscope system. Visit our Tensortek.com webpage for more videos and details. Also make sure and visit the Galleries for giga-pixel 'zoomable' images.
Affordability is the Key
So you might be wondering, how did we make this so affordable, if the big companies can't? There are three main ways we were able to cut costs. (1) Instead of using very expensive servo-motors and proprietary automated stages, we designed ours with inexpensive stepper motors and controllers using the existing stage and focus knob. This likely cuts about $20k from the cost of the 'big company' products. New advances in inexpensive micro-stepping controllers from companies like Sparkfun and Polulu have (only recently) made this possible. (2) We designed and wrote the control and imaging software ourselves, providing direct control of the motors from a standard parallel port on the computer. Moving the motors becomes as simple as a one line command (e.g. move(a:1000) means move motor 'a' clockwise 1000 steps). It's that easy. This eliminates all the expensive stepper motor controllers. (3) We designed our system to retrofit newly available microscopes that are very capable but at a fraction of the cost of the big names.
Another of our long term goals is to make this kind of technology affordable to educational and research institutions in developing countries that have highly capable and motivated people but lack of funds to buy proper equipment stops them from accomplishing great things. This is our small contribution to take them one step closer to their goals. We are working on making much more affordable scopes for those who cannot afford to spend $20k+ on a microscope (for a lot of educational institutions in developing countries that is a huge chuck of their ‘annual’ budget).
This affordability means that this can not only be used by researchers, but also by artists and DIY enthusiasts who want to explore exciting avenues that were blocked by the huge price tags.
Why we need funding
We have already developed the basic software and hardware design, and currently have three working prototypes. However, so far I have funded this all out of my own pocket. We need funding for several important improvements: (1) We want to recode the GUI in Python. I personally wrote all the VB.NET code and it is working fine right now, but it's not the best for image processing. Also, Python is open-source and we just love the basic philosopy of the language. (2) We want to develop a new super-low cost prototype for K12, developing countries, and DIYers. This means we are going to need to experiment with low cost cameras and newly available low cost scopes. (3) We need a small lab/office space. (4) If this project is funded, then I want to be working on it full-time (with the interest so far, I think I'm going to have to :). We view this kickstarter project as just the start of a much larger vision, and an exciting future where everyone wins. We realize our funding goal is high, but we need to achieve a minimum volume of systems for startup. Also, If we achieve our volume/target, we will be able to offer the system at the lowest possible price.
The pledges and rewards are designed for researchers, hobbyists, artists and supporters of collaborative work and ‘open’ idea sharing. So please come forward and help us accomplish this mega feat in the ‘micro’ world.
We have a wide range of rewards for supporting our project: cool looking 3-D lenticular magnets of 3-D images from the Mezoscope, high quality prints/posters of images of your choice from our gallery, a beautiful coffee-table style book of images (and descriptions) of our best images, and finally your own Mezoscope (camera extra, you can choose). Please view the rewards list, and very much thank you if you decide to contribute!
Final Words and Thanks!
We're super excited to bring the automated microscope you've always wanted, but until now couldn't afford. With your help & pledges, we can make this a reality!
Todd Doehring, and the Tensortek team.
visit Tensortek.com for more details and image galleries.
Support this project
- (45 days)