Tommy Johansson

The road to reliable ground truth starts before any labeling is even made. In perception software development, when training a network, the specifications of the wanted function needs to be translated to the labeling guideline, the function and its wanted behavior needs to be described in the labeling guideline to get the expected behavior from the model in the end. During annotation ambiguous situations frequently pop up, these have to be dealt with, preferably before starting the labeling work, a guideline shall also be understood and interpreted in the same way by sometimes hundreds of persons, usually with different cultural background and education.How can we achieve a solid guideline that captures all function needs and has no room for ambiguities and how can we assure that the rules/requirements in the guideline are fulfilled? This presentation will show a structured way to minimize the risk of ambiguities and inconsistency in the  guidelines and the tools available at Annotell to help you.

Since 2006 I have worked with AD/ADAS in a range of different OEM's as well as suppliers of AD/ADAS systems. In fall 2021 I joined Annotell since I realized the future of AD/ADAS will be dependent on the capabilities of deep learning and the safety argumentation around it.At Annotell I work as a Perception Expert guiding our customers how to achieve the ground truth they need with the quality they need. I'm also managing a new team called Perception Research with the goal to identify and drive Annotell’s unique position in making safe perception possible.