Building AI That Learns and Adapts: A Case Study in MRI Diagnostics
summary
Project builds a U‑Net variant fed with T1 and FLAIR plus morphometric filters (junction, extension) and trains on an initial 201 MRI series (≈200 slices each) with 201 FCD cases then a more realistic set with 30 FCDs, 150 normals and 120 other pathologies. Elastic weight consolidation and continual learning reduce catastrophic forgetting, clinical validation finds baseline detects ~70% with 54% precision and later continual learning raises specificity to about 70% while keeping sensitivity near 80%.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 3:22Focal cortical displasias (FCDs)
- 10:11AI model for FCD detection
- 23:34Continual learning matters
- 31:04Continual learning methods
- 37:49Continual learning for FCD detection
- 45:40Summary
related talks
43mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
49mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
42mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
44mBuilding Adaptive AI for MRI Diagnostics: A Continual Learning Case Study
15mBrain Tumor Detection and Classification Using Deep Learning Models
34mWhy AI in the Medical Field Both Sucks and Rocks...
34mArtificial intelligence to accelerate MRI scans - Philips & LUMC winners in the...
36mAI in Healthcare