M. Moein Shariatnia
I’m passionate about practical applications of ML in medical imaging. I’ve developed SOTA models for medical image segmentation and recently for accurate bone age estimation from hand X-rays, which could save patients from unnecessary costs and radiation exposure.
I’ve also been involved in core ML research, with two works on contrastive learning and transfer learning accepted for poster presentations at workshops in ICML’22 and ICLR’23.
Currently, I’m leading AI research projects in collaboration with the AI team at Hospital for Special Surgery (HSS) under the supervision of Dr. Ayoosh Pareek where our focus is to automate morphology parameter measurement from MSK imaging and save orthopedists the time required to manually annotate images for surgery planning and bone age assessment.
I graduated with my MD degree from Tehran University of Medical Sciences and have been doing ML research/development since 2019.
selected publications
-
Transdeeplab: Convolution-free transformer-based deeplab v3+ for medical image segmentationIn PRIME workshop, MICCAI 2022, 2022 -
Determination of Skeletal Age From Hand Radiographs Using Deep LearningThe American Journal of Sports Medicine, 2025 -
The role of pre-training data in transfer learningMRL workshop, ICLR 2023, 2023 -