Russian Researchers Unveil First AI for Infant Brain MRI Analysis

Researchers from Yandex B2B Tech, the St. Petersburg School of Economics and Management (SHADE), and the Pediatric University of St. Petersburg have introduced Russia’s first neural network for analyzing infant brain MRIs. Operating in the cloud, the model determines the ratio of gray to white matter with over 90% accuracy within minutes.

Trained on 1,500 anonymized scans and the open MICCAI Grand Challenge dataset, the system features BIBSNet for automated labeling and a ResNet with U-Net for image segmentation.

Developers highlighted the challenge of limited data and complex labeling as key hurdles. Currently in a pilot phase at a clinic, the system is accessible to doctors via a dedicated Yandex Cloud page. Plans are in place to release it as open-source to enhance research and healthcare applications.

Also read:

• Brave Uncovers Vulnerability in Perplexity’s Comet Browser
• Anthropic Develops AI Classifier to Block Weapons of Mass Destruction Requests
• Boston Dynamics Unveils Impressive New Video of Atlas Robot

Thank you!
Join us on social media!
See you!