Counting and locating high-density objects using convolutional neural network

Mauro dos Santos de Arruda

doi:10.1016/j.eswa.2022.116555

Counting and locating high-density objects using convolutional neural network

2022-06-01

10

OA

AI

M

Mauro dos Santos de Arruda

L

Lucas Prado Osco

P

Plabiany Rodrigo Acosta

D

Diogo Nunes Gonçalves

J

José Marcato

A

Ana Paula Marques Ramos

E

Edson Takashi Matsubara

Jonathan de Andrade Silva

W

Wesley Nunes Gonçalves *

DOI：10.1016/j.eswa.2022.116555

Original

Save

View PDF

Abstract

En 中文

This paper presents a Convolutional Neural Network (CNN) approach for counting and locating objects in high-density imagery. To the best of our knowledge, this is the first object counting and locating method based on a feature map enhancement combined with a multi-sigma refinement of the confidence map. The proposed method was evaluated in two counting datasets: trees and cars. For the tree dataset, our method returned a mean absolute error (MAE) of 2.05, a root-mean-squared error (RMSE) of 2.87 and a coefficient of determination (R-2) of 0.986. For the car dataset (CARPK and PUCPR+), our method was superior to state-of-the-art methods. In the these datasets, our approach achieved an MAE of 4.45 and 3.16, an RMSE of 6.18 and 4.39, and an R-2 of 0.975 and 0.999, respectively. We conclude that the proposed method is suitable for dealing with high object-density, returning a state-of-the-art performance for counting and locating objects.

Keywords:

Deep learning

Object counting

Tree counting

Car counting

Counting and locating high-density objects using convolutional neural network

Abstract

AI Summary

Journal

Organization