Title: TOPSIS method applied to the choice of the ideal object detection algorithm for an educational mobile application

Authors: André de Jesus Araújo Ramos; Marcos César da Rocha Seruffo; Roberto Célio Limão de Oliveira

Addresses: Institute of Technology, Federal University of Pará (UFPA) – University Campus of Guamá, Rua Augusto Corrêa, 01, Guamá, Belém, Pará, CEP 66075-110, Brazil ' Institute of Technology, Federal University of Pará (UFPA) – University Campus of Guamá, Rua Augusto Corrêa, 01, Guamá, Belém, Pará, CEP 66075-110, Brazil ' Institute of Technology, Federal University of Pará (UFPA) – University Campus of Guamá, Rua Augusto Corrêa, 01, Guamá, Belém, Pará, CEP 66075-110, Brazil

Abstract: This paper performs a literature review of the main methods of detection and classification of objects in real time, extracting the main characteristics that may influence their performance in smartphones. The main objective is, through the TOPSIS method, to choose the most suitable algorithm to later apply it in the task of detecting geometric shapes in streaming, from devices itself, efficiently, considering the limitations of such apparatus. This investigation contributes to making the best decision and, in subsequent works, it is intended to experimentally attest to the efficiency of such a choice in this objective. The adopted multi-criteria method indicated the YOLOv3 algorithm as the most suitable for this purpose and placed the faster-RCNN as the least ideal solution. This work brings a useful contribution to academia, as it presents a well-founded suggestion on how to popularise the use of this intelligence to object detection and classification in portable devices.

Keywords: multi-criteria; object detection algorithm; smartphone; TOPSIS; education.

DOI: 10.1504/IJMDM.2024.137075

International Journal of Management and Decision Making, 2024 Vol.23 No.2, pp.252 - 264

Received: 19 Dec 2022
Accepted: 07 Feb 2023
Published online: 01 Mar 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article