Integrating big data, video annotation and cloud based technologies for improved ADAS and Digital Mapping

The Problem

The automotive industry needs tools that can manage the extremely large volumes of data (Big Data) especially to provide support in the annotation task (ADAS, cartography market). One of the main bottlenecks in advancing in several application domains  is the lack of labelled realistic video datasets of sufficient size, complexity and coverage (comprehensiveness).

The performance of computer vision or video analysis systems is inherently restricted by the quality of the available training data. Manually collating and annotating such datasets is:

infeasible,
impractical,
slow,
inconsistent,
and excessively costly.

Cloud LSVA Solution

Cloud-LSVA will analyse and decompose each recorded scene, in order to detect and classify relevant objects and events for specific scenarios. Furthermore, the mined and annotated video metadata shall be used to train and evaluate algorithms for real-time analysis of visual and non-visual sensors in cars. We will be testing the platform in 2 scenarios:

ADAS

Analysis and annotation of petabytes of data to train and validate visual, radar and telemetry sensor data to create continuously improving ADAS algorithms for deployment in motor vehicles with possible applications in autonomous vehicles and robotics.

Digital Cartography

Street and lane level analysis and interpretation of video to automatically create new digital maps for navigation applications and provide assisted positioning (i.e. in urban canyons, underground parking structures, complex flyovers, tunnels) for deployment in motor vehicles with certain application in autonomous vehicles and robotics

Consortium

TUE dcu
ibm valeo tomtom
cea intempora ulim
ertico tass intel