Why teaching AI about symmetry could make prenatal diagnosis more reliable
Researchers at MIT, Boston Children’s Hospital and Inria introduced a deep‑learning model called E(3)-Pose that improves fetal brain MRI by tracking head motion and producing high‑resolution 3D images. The model embeds left‑right symmetry of the fetal head into its architecture, allowing it to align 2D slices taken while the fetus moves and reduce gaps in the final scan.
Key points
- E(3)-Pose embeds left‑right symmetry to align fetal MRI slices despite motion
- Model developed by MIT, Boston Children’s Hospital and Inria; presented at ECCV 2026
- Goal: faster, higher‑quality fetal brain scans to aid early diagnosis
First author Ramya Muthukrishnan, an MIT PhD student in Polina Golland’s lab, built the system with help from radiology assistant professor Borjan Gagoski and research assistant Aryn Lee, who acquired the MRI data. The work was presented at the 2026 European Conference on Computer Vision. The team says E(3)-Pose could speed up scans, make them more comfortable for pregnant mothers, and provide radiologists with clearer images for early detection of brain abnormalities. Future plans include adapting the model for second‑trimester fetuses, which move more and are currently harder to image.
Why teaching AI about symmetry could make prenatal diagnosis more reliable
MIT News on AI · 19 October 2026
Loading the full article…
This text was published by MIT News on AI and written by Alex Ouyang | Abdul Latif Jameel Clinic for Machine Learning in Health. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Nick Cammarata claims LLMs might be easier to understand · 1 src
- Opinion: 300 AI papers a day and 3.14 million models—habits to avoid drowning in AI news · 1 src
- Supply chain optimization could cut AI data center emissions, experts say · 1 src
- Study shows multi-sample network inference robustness varies with prior knowledge and data · 1 src
- Jack Clark says singularity has begun, citing AI test scores · 1 src
Comments
via GitHub Discussions