Research from the Technical University of Munich suggests artificial neural networks may be used to evaluate scans of entire mice and can display organs from bright colors, instead of in black and white.
The self-learning algorithms can be applied in biologic image processing. The University's "AIMOS" project is the work of TUM in Munich and Helmholtz Zentrum München. The algorithms were provided with the photos of mice and were positioned by the respective organs in 3D scans. The main purpose of this project was to allocate picture points to organs, such as stomach, kidneys, liver, spleen, or brain.
Because Machine learning can interpret human brain scans quicker and more reliably than humans. Computer systems are able to do equations without support from humans. The technology is expected to play an important role in basic research.
AIMOS stands for AI-based Mouse Organ Segmentation. At the heart of the program is an artificial neural network capable of learning.
"We were lucky enough to have access to several hundred image of mice from a different research project, all of which had already been interpreted by two biologists," recalls Schoppe as quoted in Science Daily.
Algorithms in the AIMOS project were conditioned on photographs of mice. The goal was to allocate 3D full-body scan imaging points to individual organs, such as the stomach, kidneys, liver, spleen, or brain. The software can then be shown the exact location and form.
With the assistance of 200 more whole-body scans of mice, the team then tested the durability of the artificial intelligence. "The result shows that self-learning algorithms are not only faster at analyzing biological image data than humans, but also more accurate," says Professor Bjoern Menze, Head of the Image-Based Biomedical Simulation Department at TranslaTUM at the Scientific University of Munich.
The new platform can be made particularly useful in basic science programs.