Browse Publications Technical Papers 2024-28-0022

Towards automation of reference data generation for ADAS/AD functions development – ALiVA framework 2024-28-0022

The advancements towards autonomous driving have propelled the need for reference/ground truth data for development & validation of various functionalities. Traditional data labelling methodologies are time consuming, skills intensive & have many drawbacks. These challenges are addressed through ALiVA (automatic lidar, image & video annotator), a semi-automated framework assisting for event detection & reference data generation through annotation/labelling of video & point-cloud data. ALiVA is capable of processing large volumes of camera & lidar sensor data. Main pillars of framework are object detection-classification models, object tracking algorithms, cognitive algorithms & annotation results review functionality. Automatic object detection functionality generates precise bounding box around the area of interest & assigns class labels to annotated objects. Object tracking algorithms tracks detected objects in video frames, provides a unique object id for each object & performs distance ranging. Unique feature offered by cognitive algorithms eliminates non-realistic objects of interests appearing in billboards or advertisements on buses/trucks. Framework also has a feature of event detection like overtaking scenarios or pedestrians/animals crossing the roads. Annotation review functionality is provided where assessment & correction of auto annotated data can be done manually. The results can be saved in universal standard file formats (txt, csv, json, openASAM). ALiVA replaces traditional annotation methodologies, resulting in reduced efforts, skilled resource requirements & time taken for annotation of large datasets ruling out the possibility of human biases, manual mistakes & inconsistencies. AliVA is validated for multiple client requirements with volume and variety of data to quantify the benefits offered. One of the differentiators is models and functionalities optimized especially for Asian road scenarios typically characterized by very high road assets density. It is platform-agnostic, adaptable to newer requirements, additions of newer event definitions for data segmentation, functioning both in cloud environments for Data as a service, and standalone desktop application.


Subscribers can view annotate, and download all of SAE's content. Learn More »

Attention: This item is not yet published. Pre-Order to be notified, via email, when it becomes available.
Members save up to 16% off list price.
Login to see discount.
Special Offer: Download multiple Technical Papers each year? TechSelect is a cost-effective subscription option to select and download 12-100 full-text Technical Papers per year. Find more information here.