PICTURE: Philipp Sodmann (left) and Matthias Griebel have developed a deep learning model that can evaluate microscopic images. view More
Photo credit: University of Würzburg
Information technology can make life easier in many areas – including research. In medicine, for example, it is still customary to evaluate microscopic images of tissue sections by hand. This is used to assess, for example, how many cancer cells are in a lymph node.
“You often sit for hours in a dark room and count the cells on an image that was taken with a fluorescence microscope. That costs an incredible amount of valuable time, ”says Philipp Sodmann, who works in heart research at the Würzburg University Hospital in Bavaria. Germany.
But now a new horizon is opening up for the life sciences: The new digital tool deepflash2 makes the analysis of microscopic images considerably easier. deepflash2 is freely available and is based on machine learning methods.
Jury emphasizes quality aspect
Matthias Griebel from the Chair of Information Systems and Business Analytics at the University of Würzburg developed the tool as part of his doctorate. The tool formed the basis for the solution developed together with the physician Philipp Sodmann for an international data science competition. The team of the two from Würzburg was successful in this competition: In May 2021, they received the innovation prize endowed with 10,000 US dollars and a gold medal from the Kaggle online platform.
The top-class jury with experts from medicine, biology and artificial intelligence (AI) attested deepflash2 a further quality: the program also recognizes ambiguities.
“In biology, not everything is black or white,” explains Matthias Griebel. It is not uncommon for researchers to doubt whether the cells they see in a tissue section are still functional. In such cases deepflash2 points out: You have to look again! In the opinion of the jury members, this makes the tool particularly innovative.
Available for free to researchers
deepflash2 is still an insider tip for researchers who deal with biological image analysis. However, Matthias Griebel would like to use the excellent results in the data science competition as an opportunity to advertise his tool.
Since it is an open source tool, other researchers can use it for free in their browser or install it on their computer. In the meantime, Griebel is already working on further improving deepflash2 with the knowledge of the competition.
deepflash2 on Github: https://matjesg.github.io/deepflash2/
Can also be used without AI knowledge
Griebel, who studied business informatics, is doing his doctorate under Professor Christoph Flath. When developing deepflash2, he attached great importance to the fact that researchers without AI expertise could use the tool without any problems.
Users from medicine and life sciences are not allowed to understand the complicated processes behind the scenes. For them, according to Griebel, the main thing is to make biographical analysis faster and at the same time more reliable. To do this, an artificial neural network with extensive data sets has to be trained intensively, says the Würzburg scientist.
Decisions are made by people
Ultimately, however, it is the people who draw a conclusion from the images. This should reassure all those who fear that artificial intelligence will decide the weal and woe of medicine in the future. Philipp Sodmann emphasizes that this is not the case and certainly will not be the case anytime soon.
Sodmann appeals to the diverse possibilities of AI to be recognized. The data science competition, for example, took place against the background of the “Human BioMolecular Atlas Program” project launched in 2016. His goal is to map and characterize every single one of the roughly 37 trillion human cells. This would not be possible without AI.
Award for the best presentation
A total of around 1,200 teams from more than 50 countries submitted solutions for the Kaggle Data Science competition. Matthias Griebel and Philipp Sodmann landed on 10th place.
“The first places were decided in a head-to-head race,” says Griebel. The presentation of the project to an international audience was also exciting for him and his colleagues. The two Würzburgers prevailed again: In addition to the gold medal and the innovation award, they also won the award for the best presentation.
Suitable for various clinical pictures
Matthias Griebel doesn’t want to do research in an ivory tower. It is important to him to develop tools that will ultimately help people. And maybe even save lives.
If microscopy images can be evaluated faster and more reliably, a diagnosis can also be made more quickly. And that applies to very different diseases. Since the deepflash2 program can be trained, it can learn, for example, to recognize different functional tissue units. With the help of machine learning, the algorithm can be taught to identify the insulin-producing cells of the pancreas on an image.
###
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of press releases sent to EurekAlert! by contributing institutions or for the use of information via the EurekAlert system.
Comments are closed.