We wanted to come up with a solution that can help people identify any magnified image regardless of their professional expertise. Initially build at HackMIT 2015.


ArisKnight Winfree Poseidon Ho David Huculak Kelly Lampotang

What it does

It takes an image of a cell sample, magnifies the image, sends it to our server, processes the image against our trained machine learning classifications, and returns if it is a match.

How We built it

Celluscope uses a lens that is attached to a smartphone's camera. The images are magnified by 1000X. The image is then uploaded to our server, and matched against our machine learning classifications. We used Clarifai's machine learning API and trained it to identify health red blood cells, malarial cells, and sickle cells. It is compared against each classification and the highest match is returned.

What's next for Celluscope

Celluscope can be used for anything, not just blood samples. We plan on increasing the size of our library for our machine learning classifications to increase the accuracy of identification. We also want to increase the number of diseases that we can identify to be large than just malaria and sickle cells.

Share this project: