| I think this project in particular is not hugely useful, the optics aren't really up to snuff to do anything other than look at small things with your eyes, it's pretty tricky to use it, and hard to take a picture with it. (Also in terms of taking a picture you are zooming the image in with the microscope lens, then zooming it back out with your cell phone lens) That being said, this project is important philosophically I think and projects like it are going to be huge. The real 'killer scopes' for me would be 1. 10-15 dollar scope with battery and camera. This is possible as raw camera chips only require 10x magnification (as opposed to the 100-600x this is providing) so it's cheap/easy to get better optics. One could imagine a low cost camera chip attached to a lens like this scope's (but a bit better) and a rasberry pi zero+battery. A scope like that would let you: (i) Set up machine learning diagnostics for malaria/other blood borne pathogens/squamous cells. Right now it takes experts to diagnose most of these diseases and those experts need good scopes. If you had thousands of these scopes in the wild and took millions of images+diagnoses, you could really fire some high powered machine learning at it to get rapid, easy diagnoses. Presumably, if you had good automated diagnostic systems (kind of pie in the sky) you could also use this for epidemiology. (ii) Sample monitoring for any kind of growth system. One could imagine a floating version of this that takes images every few minutes and keeps a count of the number of cells growing in a medium (in it's FOV), the number/type of plankton/bacteria it could see, crystal formation etc. Basically any time a researcher wants to passively monitor a process, a cheap scope like this could be useful. You could even pair a pump or something to pull in and trap your sample as needed. (iii) Water supply testing (for microbes). It's possible you could use a system like this to take a quick sample of any water supple to decide if it's potable. With a lot of people doing this you could monitor water supplies much more effectively. 2. A 30-50 dollar fluorescent microscope with camera. This is being worked on by several neuroscience labs. It's around 100-1000 dollars depending on who you believe. This could be used for: (i) Long term monitoring of cell firing in animal brains. (ii) Monitoring chemi or bioflurescenct samples in vivo long term. Could be used for levels of compounds, disease progression, drug penetration and life time. Basically this is stuff people are doing it would just make it much much cheaper and much more portable allowing longer term, higher throughput experiments. Cheap scopes are everywhere, there are a number of decent ones you can buy and hook to your standard smartphone (though most are weak because as I said before, they basically have to 'fix' your smartphone's lens). You could even make a 10 dollar USB scope yourself by cracking open a webcam and flipping its optics. But mass produced, optically good scopes with cell level resolution in everyone's hands (especially if they can easily capture those images) have the potential to make huge, really powerful databases of microscopy data. Right now computer vision in microscopy is horribly task focused, it would be cool to see what people could do with billions of diverse microscopy images. |