/anomaly-detection

Some Jupyter Notebooks with encoding models for image anomaly detection

Primary LanguageJupyter Notebook

Anomaly Detection Notebooks

This repository contains some Jupyter notebooks with examples of still-frame anomaly detection.

Table of Contents:

  • Anomaly Detection (PCA)

    • Anomaly detection using (linear) PCA on the raw RGB channels of an image

  • Anomaly Detection (JPEG-style PCA)

    • Anomay detction using linear PCA applied to the discrete cosine transform (DCT) of the channels of an image

  • NEON CAE

    • A convolutional autoencoder trained on the NEON phenocam datasets (with fake anomalies)