ConceptTTA
Test-Time Augmentation
A technique that applies multiple transformations to inputs during model inference and ensembles the prediction results. It does not change model parameters and improves robustness and accuracy by increasing prediction diversity.
Timeline (last 90 days)
Oct 19
For multi-label medical image classification tasks, a simple flat model combined with label-correlation modeling and test-time augmentation may be more effective than a carefully designed multi-stage pipeline.
Unverified50%