MH-MetroNet-A Multi-Head CNN for Passenger-Crowd Attendance Estimation.

Pier Luigi Mazzeo,Cosimo Distante,Riccardo Contino,Vito Renò,Paolo Spagnolo,Massimiliano Nitti,Ettore Stella

doi:10.3390/jimaging6070062

Pier Luigi Mazzeo, Cosimo Distante + Show 5 more

Open Access

https://doi.org/10.3390/jimaging6070062

Copy DOI

Abstract

Knowing an accurate passengers attendance estimation on each metro car contributes to the safely coordination and sorting the crowd-passenger in each metro station. In this work we propose a multi-head Convolutional Neural Network (CNN) architecture trained to infer an estimation of passenger attendance in a metro car. The proposed network architecture consists of two main parts: a convolutional backbone, which extracts features over the whole input image, and a multi-head layers able to estimate a density map, needed to predict the number of people within the crowd image. The network performance is first evaluated on publicly available crowd counting datasets, including the ShanghaiTech part_A, ShanghaiTech part_B and UCF_CC_50, and then trained and tested on our dataset acquired in subway cars in Italy. In both cases a comparison is made against the most relevant and latest state of the art crowd counting architectures, showing that our proposed MH-MetroNet architecture outperforms in terms of Mean Absolute Error (MAE) and Mean Square Error (MSE) and passenger-crowd people number prediction.

Highlights

Smart Cities are becoming real day by day, connecting the physical world to the virtual world, and, concepts such as the crowd monitoring and managing are needful
It is possible to see the performance reached by the network using different backbones. On these datasets our architecture with ResNet-152 backbone achieves the best results in terms of Mean Absolute Error (MAE) and Mean Square Error (MSE)
This paper proposed a new a multi-head Convolutional Neural Network (CNN) architecture able to perform crowd analysis estimating the density map and counting the number of people

Summary

Introduction

Smart Cities are becoming real day by day, connecting the physical world to the virtual world, and, concepts such as the crowd monitoring and managing are needful. This way crowd analysis, is becoming a hot topic in artificial intelligence because of its strong value applied in many smart cities tasks: video surveillance, public safety, urban planning, behavior understanding and so on. Real time crowd information can be usable by intelligent devices, such as smart phones and smart cameras Ullah and his team try to detect anomalies in a flow of pedestrians in such a way as to ensure their safety [1].

Methods

Results

Conclusion