MIRTO: Registration-Gated Evaluation Protocol for Brain MRI Anomaly Segmentation
What it is
MIRTO is an evaluation protocol for unsupervised anomaly detection in brain MRI that standardizes registration checks, thresholding, and statistical inference. It addresses inconsistencies in current benchmarks by explicitly measuring how evaluation choices like alignment and false-positive budgets affect performance metrics.
The problem it solves
Current evaluations of unsupervised anomaly detection rely on single scores that hide critical, unreported choices regarding image registration, thresholding, and metric aggregation, leading to inconsistent and potentially misleading performance rankings.
What you could use it for
- Benchmarking unsupervised anomaly detection methods on brain MRI datasets
- Diagnosing performance drops caused by geometric misalignment in segmentation maps
- Comparing anomaly detection models under controlled false-positive constraints
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.