%0 Conference Paper %F Oral %T Adapted learning for polarization-based car detection %+ Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes (LITIS) %+ Equipe Apprentissage (DocApp - LITIS) %+ Imagerie et Vision Artificielle [Dijon] (ImViA) %A Blin, Rachel %A Ainouz, Samia %A Canu, Stephane %A Mériaudeau, Fabrice %< avec comité de lecture %3 FOURTEENTH INTERNATIONAL CONFERENCE ON QUALITY CONTROL BY ARTIFICIAL VISION %B 14th International Conference on Quality Control by Artificial Vision %C Mulhouse, France %Y Christophe Cudel; Stéphane Bazeille; Nicolas Verrier %I SPIE %S Proceedings of SPIE %V 11172 %N UNSP 1117218 %P 98 %8 2019-05-15 %D 2019 %R 10.1117/12.2523388 %K Object detection %K Polarization imaging %K Machine Learning %K Deep Learning %Z Computer Science [cs] %Z Computer Science [cs]/Artificial Intelligence [cs.AI] %Z Physics [physics]/Physics [physics]/Optics [physics.optics]Conference papers %X Object detection in road scenes is an unavoidable task to develop autonomous vehicles and driving assistance systems. Deep neural networks have shown great performances using conventional imaging in ideal cases but they fail to properly detect objects in case of unstable scenes such as high reflections, occluded objects or small objects. Next to that, Polarized imaging, characterizing the light wave, can describe an object not only by its shape or color but also by its reflection properties. That feature is a reliable indicator of the physical nature of the object even under poor illumination or strong reflections. In this paper, we show how polarimetric images, combined with deep neural networks, contribute to enhance object detection in road scenes. Experimental results illustrate the effectiveness of the proposed framework at the end of this paper. %G English %L hal-02471097 %U https://u-bourgogne.hal.science/hal-02471097 %~ UNIV-BOURGOGNE %~ INSA-ROUEN %~ LITIS %~ COMUE-NORMANDIE %~ UNIROUEN %~ UNILEHAVRE %~ INSA-GROUPE %~ IMVIA %~ ANR