Good practices for collecting images

The success depends on the quality of the training dataset. Please consider the following points when preparing a training dataset:

General image collection advice

Specific advice for anomaly detection:

• Collecting the normal variance of the object or scene is crucial. In other words: try to collect examples of all different cases that are considered "Good", as any deviation from the training data will be classified as an anomaly.
• Anomaly detection can work with either only normal images or a combination of normal and anomalous images. If you have anomalous images, it is recommended to include them as it usually leads to higher accuracy.