Análisis del comportamiento de clientes mediante técnicas de inteligencia artificial y visión por computadora
Date
2023-10-01
Authors
Apablaza, Salvador
Fernández, Carolina Sol
Locatti, Nicolás Gabriel
Durante, Diego
Verrastro, Sebastián
Gómez, Juan Carlos
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
En el presente trabajo se desarrolla un framework de detección de personas para analizar el comportamiento de los clientes en una tienda minorista mediante el procesamiento de videos de vigilancia. Para ello, se propone un pipeline compuesto por un preprocesamiento del video, un procesamiento con un modelo basado en redes neuronales, y un postprocesamiento orientado a correcciones. Se proveen indicadores útiles para los negocios, entre ellos, un mapa de calor que muestra los espacios ocupados. El resultadoes constatado con un etiquetado manual para evaluar el rendimiento del algoritmo. Se demuestra la viabilidad de este método para comprender el comportamiento de los clientes.
In the present work, a framework for people detection is developed to analyze customer behavior in a retail store using surveillance videos. For this purpose, a three-stage pipeline consisting of video preprocessing, processing with a neural network-based model, and correction-oriented post-processing is proposed. The pipeline provides useful business indicators, including a heatmap showing occupied spaces. The results are verified through manual labeling to evaluate the algorithm's performance. The study concludes by demonstrating this method's viability for understanding customer behavior.
In the present work, a framework for people detection is developed to analyze customer behavior in a retail store using surveillance videos. For this purpose, a three-stage pipeline consisting of video preprocessing, processing with a neural network-based model, and correction-oriented post-processing is proposed. The pipeline provides useful business indicators, including a heatmap showing occupied spaces. The results are verified through manual labeling to evaluate the algorithm's performance. The study concludes by demonstrating this method's viability for understanding customer behavior.
Description
Keywords
detección de personas, visión por computadora, inteligencia artificial, mapa de calor, comportamiento de clientes, people detection, computer vision, artificial intelligence, heatmap, customer behavior
Citation
Proyecciones Vol. 21 (2)
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwised noted, this item's license is described as openAccess