CANCE detects anomalies by estimating the density of a composite feature that concatenates autoencoder latent codes with squared-error and cosine reconstruction metrics.
Imagenet: A large-scale hier- archical image database
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Anomaly Detection via Autoencoder Composite Features and NCE
CANCE detects anomalies by estimating the density of a composite feature that concatenates autoencoder latent codes with squared-error and cosine reconstruction metrics.