Menu

Software

Major KymoButler update and ImageJ plugin available now!

,

Hello everyone, we hope you are safe and following the advice regarding the current COVID-19 pandemic.

Our hearts go out to all of those who are affected by COVID-19 during this challenging time.

With many of us scientists working from home and investing our time on data analysis, we thought this would be a good opportunity to release some new KymoButler updates and tools for you.

Major KymoButler update

KymoButler has now been updated to version 1.1.1, which includes many improvements in accuracy and performance. You can find all the details in the KymoButler changelog, but what you need to know is that KymoButler performs about 3% better than before!

ImageJ plugin

Through the work of the incredible Fabrice Cordelieres (University of Bordeaux), you can now use KymoButler through an ImageJ plugin.

Try it now here!

Future updates

We are working hard on finalising our new software for the analysis of cells and nuclei, always prioritising the simplicity and ease of use of our cloud-based tools. Please stay tuned and we’ll have more updates soon!

Kymobutler-press

Our KymoButler paper is out on eLife

, ,

We are extremely happy to announce that our paper “KymoButler, a deep learning software for automated kymograph analysis” has been published on the eLife journal. You can find our press release here, and an eLife digest explaining our work for a non-specialised audience here.

But we haven’t just been constantly checking our mailboxes waiting for good news from eLife – in the past few months we have been busy making all of our software run seamlessly on the cloud, so that you can access it from anywhere and at anytime without any installation! If you work with kymographs and still haven’t tried KymoButler, give it a go here. We also updated KymoButler with some minor fixes and improved stability, have a look at KymoButler’s changelog for more details!

And more is coming! We are now further optimising the KymoButler code and planning our first neural network retraining with new data to further improve performance. We are also brewing more software that will make your image data analysis easier if you work with cell data and/or nuclei data… stay tuned for more info!

Finally, if your work requires manual annotations or quantitative data analysis, and you want to speed up and remove biases from your workflow, get in touch with us and we will find the best solution for your specific application!

Title image credit: Eva Pillai

bloglogo

Improvements to KymoButler

,

Hello everyone,

It has been less than a month since we released KymoButler and we already received great feedback that is helping us shape the future of our software.

It is with great pleasure that today we announce some big improvements to KymoButler:

The Deeper Neural Network is available now!

We optimised the Deep Neural Network from our paper to compute much faster, so we are now able to provide the full network for free on our cloud app. The new version is especially good at tracing very dense lines and the example below highlights the performance on a kymograph from actin speckle microscopy:

An input kymograph that is hard to analyse
Dense input kymograph

An image showing a kymograph analysed with the previous version of KymoButler
Shallow (old) Network

An image showing a kymograph analysed with the new version of KymoButler
Deeper (new) Network

We believe this is going to improve analysis for many of you – please try it out now at our KymoButler page!

Support for kymographs with dark and light background is available now!

We announced this a few days ago on Twitter, but we think it’s worth repeating: KymoButler now works on kymographs with dark and light backgrounds.

A kymograph with a light background
Input kymograph with a light background

The beautiful output of KymoButler, from a kymograph with a light background
Output of KymoButler from a kymograph with a light background

If you tried to analyse white background images before, please try KymoButler again now.

Bidirectional KymoButler available soon!

We are making steady progress with our bidirectional KymoButler, which will enable you to follow molecules that move both forwards and backwards in kymographs (see an example on our Software page). Please stay tuned to our blog and follow us on Twitter to receive the latest news from us!

Working together

At deepMirror.ai, we aim to bring state-of-the-art AI-based technology to quantitative data analysis that for too long has relied on manual analysis. By training Deep Neural Networks, we can teach an AI to perform as well as an expert.

If your work would benefit from this kind of approach, get in touch with us and we will find the best solution for your specific problem.

bloglogo

The free Cloud version of KymoButler is available now

, ,

We are very pleased to announce the release of the public version of our KymoButler software. We have been working for a long time on how to make kymograph analysis faster, more reliable, and pain-free.

We published our paper “KymoButler: A Deep Learning Software for Automated Kymograph Tracing and Analysis” as a preprint and decided to make our software available now by providing a free Cloud-based version of our software.

At deepMirror.ai, we aim to bring state-of-the-art AI-based technology to quantitative data analysis that has for too long relied on manual analysis. By training Deep Neural Networks, we can teach an AI to perform as well as experts.

If your work would benefit from this kind of approach, get in touch with us and we will find the best solution for your specific application.