Tracing

There are several approaches to tracing features to setup a deep learning training. Below are examples of three recommended approaches. Tracing can be done directly in the Deep Learning Trainer (fully manual, or semi-automated), in the Image Processor, or in an external application.

• In AI Session Processor
• In Image Processor
• In External Application

*Tip: The quality of model training will depend on tracing quality, so take care to trace features as accurately as possible. Tracing all features of interest is recommended, but not required.

*Tip: Training with more layers tends to produce better results. For example, consider using 3 or 4 layers, when you might have only used 2. (Example: Detecting faint pores in the presence of dark artifacts. Use at least 3 layers: pores, artifacts, and background.)