A Unified Simulation Framework for Visual and Behavioral Fidelity in Crowd Analysis

Simulation is a powerful tool to easily generate annotated data, and a highly desirable feature, especially in those domains where learning models need large training datasets. Machine learning and deep learning solutions, have proven to be extremely data-hungry and sometimes, the available real-wor...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2023-12
Main Authors: Bisagno, Niccolò, Garau, Nicola, Stefani, Antonio Luigi, Conci, Nicola
Format: Article
Language:English
Subjects:
Anomalies
Computer vision
Crowd monitoring
Deep learning
Machine learning
Pose estimation
Simulation
Simulator fidelity
Trajectory analysis
Weather
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Simulation is a powerful tool to easily generate annotated data, and a highly desirable feature, especially in those domains where learning models need large training datasets. Machine learning and deep learning solutions, have proven to be extremely data-hungry and sometimes, the available real-world data are not sufficient to effectively model the given task. Despite the initial skepticism of a portion of the scientific community, the potential of simulation has been largely confirmed in many application areas, and the recent developments in terms of rendering and virtualization engines, have shown a good ability also in representing complex scenes. This includes environmental factors, such as weather conditions and surface reflectance, as well as human-related events, like human actions and behaviors. We present a human crowd simulator, called UniCrowd, and its associated validation pipeline. We show how the simulator can generate annotated data, suitable for computer vision tasks, in particular for detection and segmentation, as well as the related applications, as crowd counting, human pose estimation, trajectory analysis and prediction, and anomaly detection.
ISSN:2331-8422