Accurate pedestrian localization in overhead depth images via Height-Augmented HOG

2018/05/31 by Werner Kroneman, Alessandro Corbetta, Federico Toschi · 4 citations
Computer Science · Engineering · #Advanced Neural Network Applications #Feature (linguistics) #Histogram #Histogram of oriented gradients #Overhead (engineering) #Pattern recognition (psychology) #Pedestrian #Pedestrian detection #Robotics and Sensor-Based Localization #Scalability #Video Surveillance and Tracking Methods #cs.CV

paper · pdf · doi:10.17815/cd.2020.30

published in Collective Dynamics 5 · 8 pages

arxiv created 2018/05/31 · openalex created_date 2018/06/13 · openalex publication_date 2020/03/27 · arxiv updated 2021/02/17 · openalex updated_date 2026/08/05

Abstract

We tackle the challenge of reliably and automatically localizing pedestrians in real-life conditions through overhead depth imaging at unprecedented high-density conditions. Leveraging upon a combination of Histogram of Oriented Gradients-like feature descriptors, neural networks, data augmentation and custom data annotation strategies, this work contributes a robust and scalable machine learning-based localization algorithm, which delivers near-human localization performance in real-time, even with local pedestrian density of about 3 ped/m2, a case in which most stateof- the art algorithms degrade significantly in performance.

Citations

• Multi-column deep neural networks for image classification
• Weakly supervised training of deep convolutional neural networks for overhead pedestrian localization in depth fields
• The Effectiveness of Data Augmentation in Image Classification using Deep Learning
• A large-scale real-life crowd steering experiment via arrow-like stimuli