Use Case: Generating AI training and test data for Synset

Generating training and test data for AI and deep learning applications has challenges and advantages. If done right, it provides a source for data that is highly controllable and can be extended with minimum human effort. This use case implements a data generation pipeline in OCTAS® for mobility applications that enables parameterizable, stochastic variation.

Two different machine learning tasks are considered: Traffic sign recognition and vehicle make and model recognition (VMMR). While each has very distinct challenges, they both leverage many of the same parts of OCTAS®. Therefore, we describe them together.

Goal