Title :
Validation and benchmarking for pedestrian video detection based on a sensors simulation platform
Author :
Bossu, Jérémie ; Gruyer, Dominique ; Smal, Jean Christophe ; Blosseville, Jean Marc
Author_Institution :
LEMCO (INRETS), Versailles-Satory, France
Abstract :
The evaluation stage is an important part in the validation of ADAS robustness. Moreover, the control and the repetitiveness of the experimentations were very difficult to conduct on real road due to safety reasons. Moreover, the lack of data/sensors or the complexity of the experiment are often very penalizing for a correct and exhaustive evaluation. It is for these reasons that LIVIC launched the development of a software simulation architecture (SIVTC), to support its research activities on ADAS. The use of such a simulation platform can provide both simulated sensors data and a ground truth reference for the validation stages.
Keywords :
computational complexity; computer vision; image sensors; object detection; traffic engineering computing; video signal processing; ADAS robustness; SIVTC; exhaustive evaluation; pedestrian video detection; sensors simulation platform; software simulation architecture; Cameras; Detection algorithms; Intelligent vehicles; Mathematical model; Protocols; Road safety; Testing; USA Councils; Vehicle dynamics; Virtual prototyping;
Conference_Titel :
Intelligent Vehicles Symposium (IV), 2010 IEEE
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4244-7866-8
DOI :
10.1109/IVS.2010.5548031
