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.
Original languageEnglish
Title of host publicationEleventh International Conference on Machine Vision, ICMV 2018
EditorsDmitry P. Nikolaev, Antanas Verikas, Petia Radeva, Jianhong Zhou
Number of pages8
Volume1104101
PublisherSPIE - International Society for Optical Engineering
Publication date15 Mar 2019
Article number110410I
ISBN (Electronic)9781510627482
DOIs
Publication statusPublished - 15 Mar 2019
EventThe 11th International Conference on Machine Vision - Munich, Germany
Duration: 1 Nov 20183 Nov 2018

Conference

ConferenceThe 11th International Conference on Machine Vision
Country/TerritoryGermany
CityMunich
Period01/11/201803/11/2018
SeriesProceedings of SPIE, the International Society for Optical Engineering
Volume11041
ISSN1605-7422

Fingerprint

Dive into the research topics of 'Occlusion-aware pedestrian detection'. Together they form a unique fingerprint.

Cite this