ZeroCostDL4Mic Democratizes Deep Learning for Microscopy Imagery — and Runs Right in Your Browser
Based on Google Collab, ZeroCostDL4Mic is designed to allow anyone access to powerful DL tools for microscopy.
A team of scientists has released a deep learning platform for microscopy, capable of running on any machine with a web browser — and it's entirely free and open source: ZeroCostDL4Mic.
"The key novelty is that ZeroCostDL4Mic runs in the cloud for free and does not require users to have any coding experience or advanced computational skills," Senior Researcher Guillaume Jacquemet explains of the platform he and his team have released. "Effectively, it runs on any computer that has a web browser."
"We believe that ZeroCostDL4Mic will acts as a 'gateway drug' for AI, luring users to explore these new technologies that will transform biomedical research and diagnostics in the decades to come."
Digital microscopy, in which the microscope either has a digital camera integrated or is designed to be connected to a digital camera, is nothing new — but actually processing the hundreds of thousands of images each sample can generate takes work. That's where ZeroCostDL4Mic comes into play, offering self-explanatory notebooks for Google Colab tied together with an easy-to-use graphical user interface.
"By bringing previously validated methods into a streamlined format that allows easy, cost-free access and customised DL use for microscopy data," the team concludes, "we believe that ZeroCostDL4Mic provides an important step towards broadening the use of DL approaches beyond the community of computer scientists to the biology laboratories that generate the imaging data."
"We hope to make DL available to all researchers regardless of their laboratory’s scale and means. We believe that this democratization will contribute to the acceptance and validation of DL methods in biomedical research."
A paper on ZeroCostDL4Mic has been published under open-access terms in the journal Nature Communications; a video demonstration using ZeroCostDL4Mic to detect and follow cancer cells as well as boost the fidelity of various microscopy images can be found on YouTube. The source code, meanwhile, is available on GitHub under the permissive MIT license.