Beyond Medical Image Segmentation. The Road Towards Clinical Insights
summary
A multi-stage pipeline processes CT scans for abdominal aortic aneurysm planning using sub-organ multilabel segmentation and a skeletonization-based loss, running inference on cloud GPU inference and producing segmentations that save about 20 minutes per case. The company assembles roughly 3k annotated CT scans and builds a similarity system based on centerline alignment, geometric features and semantic aortic segments with cosine distances, evaluated internally and used for treatment planning.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:57Cydar Ev Maps: Fully automatic Image fusion
- 2:30Medical Image segmentation
- 3:01Deep learning-based image segmentation
- 7:59What can we do about it?
- 8:52Centerline Dice
- 10:37Blocks sampling
- 12:33Outline
- 14:23Endovascular aneurysm repair
- 16:00What affects outcomes?
- 17:17How is experience disseminated?
- 20:50Our experiment
- 23:34Problem: dataset size
- 25:42Geometrical features
- 29:01Summary
related talks
11mInterview
18mCancer Screening using Deep Learning
34mWhy AI in the Medical Field Both Sucks and Rocks...
43mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
49mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
29mHow AI Is Transforming Medical Imaging
47mThe whole truth about using deep learning in medical imaging
42mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics