Augmenting an autoencoder's latent code with k-nearest-neighbor context from the same batch improves unsupervised brain-MRI anomaly detection (AUC 0.90, AP 0.78 vs 0.84/0.62 baseline).
Unsupervised Anomaly Detection in 3D Brain FDG PET: A Benchmark of 17 V AE-Based Approaches
1 Pith paper cite this work, alongside 8 external citations. Polarity classification is still indexing.
1
Pith paper citing it
8
external citations · OpenAlex
fields
q-bio.QM 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
In-batch Relational Features Enhance Precision in An Unsupervised Medical Anomaly Detection Task
Augmenting an autoencoder's latent code with k-nearest-neighbor context from the same batch improves unsupervised brain-MRI anomaly detection (AUC 0.90, AP 0.78 vs 0.84/0.62 baseline).