ikosa ai image

DoItYourself: deep learning applications, specific for your needs

Written by:

Siegfried Schwarz, MSc

Certainly, you remember when we highlighted in our blog “Angiogenesis and Artificial Intelligence, a long-overdue match” how much you and science, in general, would benefit from an intuitive tool that allows you to build robust Deep Learning applications yourself. Especially so, if the tool would not require any prior knowledge in computer science or programming for operation. Followed by us telling you to stop daydreaming.

So, what if we told you that we have been working extensively on a tool that allows you to do exactly that?

It has always been our mission to support life science research at our best. That is our driving motivation for creating a brand new tool that will be released within the next couple of weeks. It will provide functionalities to create, train and execute your own image analysis applications, all by yourselves. 

We have lately been putting a lot of thought and effort into creating such a tool that is fast, accurate and robust enough to support your individual research needs. However, most of our efforts were concentrated on designing it in a way that literally NO prior programming experience is required - that is our promise!

In other words, in the future, you will simply upload the images you want to be analyzed to our platform. Then, you will have to prepare your image dataset by marking the regions of interest - this step is no more difficult than doodling in paint, believe us. Before the training of your Deep Learning application starts, we will give you hints on how many images you need to annotate to achieve reproducible and accurate results.

As soon as you kick off the training you could go ahead and grab a coffee, have a stimulating conversation with your team or just continue summarizing your latest findings. After some time the training finishes and now you are almost ready totry out your own, brand new Deep Learning application! Test your creation with images that have not been part of the initial training and then you are done.

IKOSA AI Algorithm Training workflow

The step-by-step process of creating Deep Learning applications with IKOSA AI

Please be patient with us, as this inspiring development requires a lot of testing and improving to keep up our promise. Stay tuned for more information on IKOSA AI

PS: If you already have a specific Deep Learning application in mind that you feel cannot wait, submit your assay idea