KI: Neue Arbeitsgruppe AMBIOM

New AMBIOM Group: Jianxu Chen Wants to Develop „Eyes & Brains“

Dortmund, 20th September 2021

Medical imaging generates enormous amounts of data that can neither be managed efficiently nor evaluated thoroughly without artificial intelligence (AI). Therefore, ISAS has expanded its Biospectroscopy department (headed by Prof Dr Matthias Gunzer) with a new working group. The junior research group AMBIOM - Analysis of Microscopic BIOMedical Images, which is funded by the Federal Ministry of Education and Research, aims to enable a high analytical throughput of microscope images. Led by Dr Jianxu Chen, a renowned expert in biomedical image analysis and deep learning, the working group plans to develop algorithms and methods (open source) by 2026 that will allow countless image data worldwide to be analysed automatically, quickly and economically. “My passion is to build ‘eyes and brains’ for computers to understand image data in medical or biological studies”, says Chen. In the future, for example, the work at AMBIOM should allow broad-based new studies on the development of diseases and their consequences at the level of entire organs and organ systems. Furthermore, the AI analysis methods developed at ISAS will support doctors in making diagnostic and therapeutic decisions.

In his research, Chen has focused on developing scalable AI-based biomedical image analysis algorithms, especially for large 3D microscopy image data. Before joining ISAS, the 33-year-old worked for four years at the Allen Institute for Cell Science in Seattle. He recently moved to Dortmund with his family of four for his new job.

Dr. Jianxu Chen hat seine Forschungsarbeit am ISAS aufgenommen. — Dr Jianxu Chen has started his research at ISAS.

© ISAS

About ISAS

The Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. develops efficient and cost-effective analytical methods for health research. Thus, it contributes to the improvement of the prevention, early diagnosis and therapy of diseases. Overall, the institute strives to advance precision medicine by combining knowledge from different fields such as chemistry, biology, pharmacology, physics and computer science. The institute works closely with universities in Germany and abroad, for instance through joint appointments. It also cooperates with national and international partners from science and industry. ISAS was founded in 1952 and has around 200 employees from 20 nations.

Downloads

Further press releases

7th February 2025

The Less the Better

If neutrophil granulocytes migrate into injured tissue, for example into the brain after a stroke, they can promote chronic inflammation and cause long-term damage. In order to analyze the functional dynamics of these immune cells, millions of them are needed. A group of researchers recently developed a method that enables mass spectrometric analysis with just 1,000 neutrophils.

20th January 2025

Prolonged Health with a Compound Found in Mushrooms?

Is an active ingredient from shiitake or oyster mushrooms able to prolong ageing health? An international team of researchers led by Dr habil. Miloš Filipović has unravelled the mechanism of action of ergothioneine. The results show how the substance improves the health of ageing animals. A human study is currently being planned.

Muskulöse Ratten (Gezeichnet im Asterix-Stil) stehen um einen mit Pilzen gefüllten Zaubertranktopf herum.

23rd April 2024

A cause of immunodeficiency identified

After a stroke or heart attack, the immune system of those affected is often impaired. This can lead to life-threatening infections. A team of researchers from the Faculty of Medicine at the University of Duisburg-Essen, Essen University Hospital and ISAS, has now discovered a previously unknown cause for the connection between stroke, heart attack and immunodeficiency. But that's not all: the team has also identified a new treatment approach.

Fluoreszenz-Ultramikroskopische Aufnahme aus dem speziellen Darmgewebe einer Maus, das besonders viele Ig-produzierende Plasmazellen enthält.

24th January 2024

EfficientBioAI: New Open-Source Software Makes AI Models Lighter and Greener

Artificial intelligence has become an integral part of research. However, the better and more complex the models become, the higher their energy consumption. Researchers at ISAS and Peking University have therefore developed ready-to-use and open software that compresses existing bioimaging AI models. With the help of the new toolbox, scientists can now run their models faster and with significantly lower energy consumption.

Die Abbildung zeigt die 3D-Semantiksegmentierung von Osteozyten in Mäuseknochen vor und nach der Kompression des Modells mir EfficientBioAI Die Bilder wurden mittels Lichtblatt-Fluoreszenzmikroskopie aufgenommen.

11th December 2023

ComplexEye & AI Enable Faster Migration Analysis of Immune Cells

How and where immune cells such as neutrophil granulocytes migrate, for instance whether they infiltrate tumours, is crucial for cancer patients. They could benefit from drugs that prevent this migration. Until now, the migration behaviour of immune cells could only be investigated using conventional video microscopy. However, researchers at the University of Duisburg-Essen and ISAS have now developed the ComplexEye. With their new microscope for the high-throughput analysis of drug substances, they are now able to analyse significantly higher numbers of neutrophils for their migration behaviour.

Vergleich ComplexEye-Objektiv und 28,5-mm-Objektiv eines konventionellen Mikroskops.

11th November 2022

Chan Zuckerberg Initiative Funds Two Projects from Dortmund

To extend the image analysis platform napari, Chan Zuckerberg Initiative is funding two projects of the ISAS research groups AMBIOM - Analysis of Microscopic BIOMedical Images and Spatial Metabolomics.

Die linke Abbildung zeigt eine Mikroskop-Aufnahme von Tumor-Zellen. Auf der rechten Seite ist die Segmentierung mittels gängiger Computerprogramme zu sehen. Sobald die Zellen dicht nebeneinander liegen oder überlappen (s. blaue Markierung) verschlechtert sich die Segmentierung. Das vollautomatische Tracking führt daher im Ergebnis zu Ungenauigkeiten.