Machine Learning (ML) techniques enable automated driving. The techniques help a vehicle perceive its environment and process the data it collects. To ensure reliable functioning, ML algorithms require high-quality annotated data. However, the large volume of data and ambiguous labeling prevent the data from being reused and shared, and the results being reproduced. Besides open databases and the datasets offered by data providers have their taxonomies, formats, and data models, making it challenging to share and reuse the data.
Unambiguous labeling of data is critical for the safety of autonomous driving systems. However, it is often associated with enormous effort and considerable costs. The Association for Standardization of Automation and Measuring Systems (ASAM) OpenLABEL is the first standard to provide a solution.
ASAM OpenLABEL standard specifies a data model and format for structuring and organizing different types of sensor data into labels. In addition, it defines a set of standardized tags and a data model to categorize and organize scenarios.
To guarantee an unambiguous designation of labels, tags, and other description elements, ontologies can also be used along with ASAM OpenLABEL. Although the organization recommends the use of ASAM OpenXOntology, the standard can also be used with other ontologies and taxonomies.
An international working group of experts from 22 ASAM member companies developed OpenLABEL in line with the requirements and expectations of the industry. Potential users of the standard include specialists in machine learning, perception and/or computer vision, data labeling, test processes, systems, validation and verification of ADAS and AV, functional safety, and simulation.
Click here to read the original article published by Automotive Testing Technology International.
Please give your feedback on this article or share a similar story for publishing by clicking here.
