The demo analyzes a dataset of 3,823 images of handwritten digits where each image is 8 by 8 pixels. The demo program presented in this article uses image data, but the autoencoder anomaly detection ...

Autoencoder: A neural network architecture designed to learn compressed representations of data, ... Anomaly Detection in Images Using Deep Learning Techniques Publication Trend.

The demo program presented in this article uses image data, but the autoencoder anomaly detection technique can work with any type of data. The demo begins by creating a Dataset object that stores the ...