%0 Chart or Table
%A Abubaker, Mohamad
%A Alsadder, Zubayda
%A Abdelhaq, Hamed
%A Boltes, Maik
%A Alia, Ahmed
%T RPEE-Heads: A Benchmark for Pedestrian Head Detection in Crowd Videos
%C Jülich
%I Forschungszentrum Jülich
%M FZJ-2024-07226
%D 2024
%X RPEE-Heads (Railway Platforms and Event Entrances-Heads) is a new benchmark for pedestrian head detection in crowded environments. It focuses on railway platforms and event entrances, where risks frequently arise. The benchmark aims to improve pedestrian head detection at railway platforms and event entrances, helping to develop accurate deep learning models for several crowd safety applications. It includes: 1) A dataset comprising 109913 head annotations across 1886 images, with an average of approximately 56.2 annotated heads per image.  2) An empirical comparative analysis of eight state-of-the-art deep learning algorithms for head detection was conducted across several publicly available image datasets and the newly introduced RPEE-Heads dataset. 3) An empirical study on head size’s impact on detection algorithms’ performance.
%F PUB:(DE-HGF)32
%9 Dataset
%R 10.34735/PED.2024.2
%U https://juser.fz-juelich.de/record/1034460