Real Time Object Detection And Tracking

EasyChair Preprint 223

Abstract

The present work proposes a real-time detection and tracking system to be implemented in commercial areas. The purpose is to gather and make sense of costumer behavior data extracted from surveillance footage (available from ceiling cameras) in order supply retailers with a set of analytics, management and planning tools to help them perform tasks such as planning demand and supply chains and organizing product placement on shelfs. To achieve this goal, deep learning techniques are used, which have been yielding outstanding results in computer vision problems in recent years.

Keyphrases: computer vision, deep learning, object detection, object tracking

@booklet{EasyChair:223,
  author    = {Dária Baikova and Rui Maia and Pedro Santos and João Ferreira and Joao Oliveira},
  title     = {Real Time Object Detection And Tracking},
  doi       = {10.29007/zl3j},
  howpublished = {EasyChair Preprint 223},
  year      = {EasyChair, 2018}}