Combining Video and Wireless Signals for Enhanced Audience Analysis

Abstract

We present a system for audience engagement measurement which combines wireless and vision-based detection techniques. The system is able to detect the position and the movements of the audience during a live event with rapidly varying illumination. At the heart of the paper is an approach to use a wireless-based person detection and tracking system to guide the preprocessing of the frames which are fed to the CNN performing person analysis. We show that the hybrid system performs better than standard vision-based approaches and can be successfully deployed in environments with challenging illumination conditions.

BIB_text

@Article {
title = {Combining Video and Wireless Signals for Enhanced Audience Analysis},
keywds = {
Multi-modal Analysis, Person Analysis, Wireless Detection, Computer Vision, Audience Engagement
}
abstract = {

We present a system for audience engagement measurement which combines wireless and vision-based detection techniques. The system is able to detect the position and the movements of the audience during a live event with rapidly varying illumination. At the heart of the paper is an approach to use a wireless-based person detection and tracking system to guide the preprocessing of the frames which are fed to the CNN performing person analysis. We show that the hybrid system performs better than standard vision-based approaches and can be successfully deployed in environments with challenging illumination conditions.


}
isbn = {978-989758395-7},
date = {2020-02-22},
}
Vicomtech

Parque Científico y Tecnológico de Gipuzkoa,
Paseo Mikeletegi 57,
20009 Donostia / San Sebastián (España)

+(34) 943 309 230

Edificio Ensanche,
Zabalgune Plaza 11,
48009 Bilbao (España)

close overlay