139 Downloads (Pure)

Abstract

Failure in pedestrian detection systems can be extremely crucial, specifically in driverless driving. In this paper, failures in pedestrian detectors are refined by re-evaluating the results of state of the art pedestrian detection systems, via a fully convolutional neural network. The network is trained on a number of datasets which include a custom designed occluded pedestrian dataset to address the problem of occlusion. Results show that when applying the proposed network, detectors can not only maintain their state of the art performance, but they even decrease average false positives rate per image, especially in the case where pedestrians are occluded.
OriginalsprogEngelsk
TitelEleventh International Conference on Machine Vision, ICMV 2018
RedaktørerDmitry P. Nikolaev, Antanas Verikas, Petia Radeva, Jianhong Zhou
Antal sider8
Vol/bind1104101
ForlagSPIE - International Society for Optical Engineering
Publikationsdato15 mar. 2019
Artikelnummer110410I
ISBN (Elektronisk)9781510627482
DOI
StatusUdgivet - 15 mar. 2019
BegivenhedThe 11th International Conference on Machine Vision - Munich, Tyskland
Varighed: 1 nov. 20183 nov. 2018

Konference

KonferenceThe 11th International Conference on Machine Vision
Land/OmrådeTyskland
ByMunich
Periode01/11/201803/11/2018
NavnProceedings of SPIE, the International Society for Optical Engineering
Vol/bind11041
ISSN1605-7422

Fingeraftryk

Dyk ned i forskningsemnerne om 'Occlusion-aware pedestrian detection'. Sammen danner de et unikt fingeraftryk.

Citationsformater