Implementaion of deep learning models in Java-based image analysis software tools

Milan, Italy

Event dates: 6-10 February 2023 (5 days)

Event location: Milan, Italy

Venue: Human Technopole

Event overview

BioImage Model Zoo, which provides a growing number of deep learning models for bioimaging applications constitutes a serious milestone in terms of bringing Artificial Intelligence (AI) to the life sciences. Being able to run models stored in the BioImage Model Zoo in Java-based image processing programs such as DeepImageJ, Labkit (Imaris), Icy, QuPath, etc. would enable users of these image analysis applications to deploy AI to analyze small and large image datasets. Recently, an imglib2 API was developed around the deep-java library, which makes it possible to run Model Zoo models in Java. However, more work is required to make the end-user experience better. For example, code that automatically downloads the models and corresponding jars for the respective deep learning frame-work and operating system need to be coded. Connectors of the core imglib2 API to the various front-ends, and the imglib2 core API itself need to be improved.

This hackathon which will take place at Human Technopole research institute aims to bring bioimage analysis enthusiasts with a strong knowledge in Java to integrate deep learning models in Java-based image analysis tools.

This event is limited to 15 participants and will take place in-person only.

Please note that this event will only go ahead if there is a sufficient number of confirmed participants. Successful applicants will be notified in due time.

Event format

This event will take place in-person only at Human Technopole institute in Milan, Italy

Attendance fee

This event is free of charge

Prerequisites for application

  • Experience and proven track record of integrating bioimage analysis models into Java-based open source image processing software

Application form

Please complete the application form hosted on SurveyMonkey:

https://www.surveymonkey.com/r/deep_learning_hackathon

Deadline for applications: 19 December 2022

Contact person

For any questions, please contact Gleb Grebnev at the following address: gleb.grebnev at embl.de