Alibi Detect
lastmod 2021-03-07created 2021-03-072 edits
Algorithms for outlier and adversarial instance detection, concept drift, and metrics.
Alibi Detect is an open source Python library focused on outlier, adversarial, and drift detection. The package aims to cover both online and offline detectors for tabular data, text, images, and time series. The outlier detection methods should allow the user to identify global, contextual, and collective outliers.
Surface Crack Detection
- Youtube - AI that finds surface defects - Outlier Detection
- Surface Crack Detection | Kaggle
- Surface Crack Outlier Detection - Colaboratory
Surface_Crack_Outlier_Detection.png
Shall we build a practical AI model that finds the cracked parts of a concrete surface using alibi-detect's VAE? This time, instead of a traditional classifier, we will use an Unsupervised learning approach that detects defects by training only on normal images.