Estimation of crowd density applying wavelet transform and machine learning

  title={Estimation of crowd density applying wavelet transform and machine learning},
  author={Koki Nagao and Daichi Yanagisawa and Katsuhiro Nishinari},

Crowd density estimation based on classification activation map and patch density level

A network named Patch Scale Discriminant Regression Network (PSDR) and a person classification activation map (CAM) method, which provides person location information and guides the generation of the entire density map in the final stage, which performs better than state-of-the-art methods.

Estimation of pedestrian crowds' properties using commercial tablets and smartphones

The proposed approach uses angular velocity and acceleration measurements obtained from inertial sensors to judge amount of body motion and therefore estimate walking velocity and density and may help getting a general picture of pedestrian crowds' condition over large areas with greater employment flexibility.

Crowd density classification method based on pixels and texture features

An adaptive crowd density classification method based on pixels and texture features which can reach to 98.2%, whose indexes of MAE and MSE also outperform the most existing methods.

Multiview-Fusion-Based Crowd Density Estimation Method for Dense Crowd

This method is applied to the crowd gathering safety situation assessment of Suzhou city life fountain square, and achieves good results, which provides theoretical support for the safety control of crowd gathering place based on the Internet of things.

A convolutional neural‐network‐based pedestrian counting model for various crowded scenes

A new network model making use of stacked multicolumn convolutional neural networks (CNNs) for pedestrian counting with considerable advantages in pedestrian counting tasks compared to other state‐of‐the‐art models is proposed and has an improvement effect for the training process.

RF-Driven Crowd-Size Classification via Machine Learning

In this letter, we propose a machine learning solution for crowd-size classification in an indoor environment. Narrow-band radio frequency signals are used to identify a pattern according to the

Overcrowding Detection Based on Crowd-Gathering Pattern Model

Different cases of crowd gathering based on Edward Hall’s personal space theory are discussed and a novel crowd gathering pattern model is constructed and a modified multi-column convolutional neural network is proposed for extracting the overcrowding.

Analysis and modelling of macroscopic and microscopic dynamics of a pedestrian cross-flow

In this work we investigate the behaviour of a human crowd in a cross-flow. In the first part of our work we analyse the results of a set of controlled experiments in which subjects were divided into



Collecting pedestrian trajectories

Empirical study of crowd behavior during a real mass event

A new method based on a flow field visualization algorithm and an analytical model of a shock wave in a crowd is established to theoretically investigate the stop-and-go wave, and the model can be used to explain the measurement results.

Improvement of pedestrian flow by slow rhythm.

We have developed a simple model for pedestrians by dividing walking velocity into two parts, which are step size and pace of walking (number of steps per unit time). Theoretical analysis on pace

Pedestrians rotation measurement in bidirectional streams

This study presents an experimental measurement of pedestrians' body rotation in bidirectional streams. A mock-up corridor monitored using a camera placed on azimuthal position is used to study

Dynamics of crowd disasters: an empirical study.

Video recordings of the crowd disaster in Mina/Makkah during the Hajj in 1426H on 12 January 2006 are analyzed and reveal two subsequent, sudden transitions from laminar to stop-and-go and "turbulent" flows, which question many previous simulation models.

Pedestrian speed/flow relationships for walking facilities in Hong Kong

This paper reports on the findings of pedestrian flow characteristics for different types of walking facilities in Hong Kong. Surveys were conducted during peak hours at indoor and outdoor walkways,

Empirical analysis of the lane formation process in bidirectional pedestrian flow.

Examining the flow formed by two groups of people walking toward each other in a mock corridor shows that balanced bidirectional flow becomes the most stable configuration after lanes are formed, but the lane creation process requires pedestrians to laterally move to a largest extent compared to low flow-ratio configurations.

Generalized centrifugal-force model for pedestrian dynamics.

A spatially continuous force-based model for simulating pedestrian dynamics is introduced which includes an elliptical volume exclusion of pedestrians and shows good agreement with empirical data obtained in controlled experiments.