Skip to Main content Skip to Navigation
Journal articles

Dataset Optimization for Real-Time Pedestrian Detection

Remi Trichet 1 Francois Bremond 1
1 STARS - Spatio-Temporal Activity Recognition Systems
CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : This paper tackles data selection for training set generation in the context of near real-time pedestrian detection through the introduction of a training methodology: FairTrain. After highlighting the impact of poorly chosen data on detector performance, we will introduce a new data selection technique utilizing the expectation-maximization algorithm for data weighting. FairTrain also features a version of the cascade-of-rejectors enhanced with data selection principles. Experiments on the INRIA and PETS2009 datasets prove that, when ne trained, a simple HoG- based detector can perform on par with most of its near real-time competitors.
Document type :
Journal articles
Complete list of metadata

Cited literature [70 references]  Display  Hide  Download

https://hal.inria.fr/hal-01566517
Contributor : Remi Trichet <>
Submitted on : Tuesday, August 1, 2017 - 2:04:02 PM
Last modification on : Tuesday, July 21, 2020 - 9:26:05 AM

File

RR - 9084.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Remi Trichet, Francois Bremond. Dataset Optimization for Real-Time Pedestrian Detection. IEEE Access, IEEE, 2017, pp.15. ⟨10.1109/ACCESS.2017.2788058⟩. ⟨hal-01566517⟩

Share

Metrics

Record views

434

Files downloads

1138