A New Loss Function for Simultaneous Object Localization and Classification
Robots play a pivotal role in the manufacturing industry. This has led to the development of computer vision. Since AlexNet won ILSVRC, convolutional neural networks (CNNs) have achieved state-of-the-art status in this area. In this work, a novel method is proposed to simultaneously detect and predi...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-03-01
|
Series: | Mathematics |
Subjects: | |
Online Access: | https://www.mdpi.com/2227-7390/11/5/1205 |
_version_ | 1797614864349265920 |
---|---|
author | Ander Sanchez-Chica Beñat Ugartemendia-Telleria Ekaitz Zulueta Unai Fernandez-Gamiz Javier Maria Gomez-Hidalgo |
author_facet | Ander Sanchez-Chica Beñat Ugartemendia-Telleria Ekaitz Zulueta Unai Fernandez-Gamiz Javier Maria Gomez-Hidalgo |
author_sort | Ander Sanchez-Chica |
collection | DOAJ |
description | Robots play a pivotal role in the manufacturing industry. This has led to the development of computer vision. Since AlexNet won ILSVRC, convolutional neural networks (CNNs) have achieved state-of-the-art status in this area. In this work, a novel method is proposed to simultaneously detect and predict the localization of objects using a custom loop method and a CNN, performing two of the most important tasks in computer vision with a single method. Two different loss functions are proposed to evaluate the method and compare the results. The obtained results show that the network is able to perform both tasks accurately, classifying images correctly and locating objects precisely. Regarding the loss functions, when the target classification values are computed, the network performs better in the localization task. Following this work, improvements are expected to be made in the localization task of networks by refining the training processes of the networks and loss functions. |
first_indexed | 2024-03-11T07:17:22Z |
format | Article |
id | doaj.art-2369e7ff71a542208ba0765882fb479d |
institution | Directory Open Access Journal |
issn | 2227-7390 |
language | English |
last_indexed | 2024-03-11T07:17:22Z |
publishDate | 2023-03-01 |
publisher | MDPI AG |
record_format | Article |
series | Mathematics |
spelling | doaj.art-2369e7ff71a542208ba0765882fb479d2023-11-17T08:09:39ZengMDPI AGMathematics2227-73902023-03-01115120510.3390/math11051205A New Loss Function for Simultaneous Object Localization and ClassificationAnder Sanchez-Chica0Beñat Ugartemendia-Telleria1Ekaitz Zulueta2Unai Fernandez-Gamiz3Javier Maria Gomez-Hidalgo4System Engineering and Automation Control Department, University of the Basque Country (UPV/EHU), Nieves Cano 12, 01006 Vitoria-Gasteiz, SpainSystem Engineering and Automation Control Department, University of the Basque Country (UPV/EHU), Nieves Cano 12, 01006 Vitoria-Gasteiz, SpainSystem Engineering and Automation Control Department, University of the Basque Country (UPV/EHU), Nieves Cano 12, 01006 Vitoria-Gasteiz, SpainDepartment of Nuclear and Fluid Mechanics, University of the Basque Country (UPV/EHU), Nieves Cano 12, 01006 Vitoria-Gasteiz, SpainMERCEDES BENZ España, Las arenas 1, 10152 Vitoria-Gasteiz, SpainRobots play a pivotal role in the manufacturing industry. This has led to the development of computer vision. Since AlexNet won ILSVRC, convolutional neural networks (CNNs) have achieved state-of-the-art status in this area. In this work, a novel method is proposed to simultaneously detect and predict the localization of objects using a custom loop method and a CNN, performing two of the most important tasks in computer vision with a single method. Two different loss functions are proposed to evaluate the method and compare the results. The obtained results show that the network is able to perform both tasks accurately, classifying images correctly and locating objects precisely. Regarding the loss functions, when the target classification values are computed, the network performs better in the localization task. Following this work, improvements are expected to be made in the localization task of networks by refining the training processes of the networks and loss functions.https://www.mdpi.com/2227-7390/11/5/1205image classificationobject detectiondeep learningdeep convolutional neural networkscomputer visioncustom training loop |
spellingShingle | Ander Sanchez-Chica Beñat Ugartemendia-Telleria Ekaitz Zulueta Unai Fernandez-Gamiz Javier Maria Gomez-Hidalgo A New Loss Function for Simultaneous Object Localization and Classification Mathematics image classification object detection deep learning deep convolutional neural networks computer vision custom training loop |
title | A New Loss Function for Simultaneous Object Localization and Classification |
title_full | A New Loss Function for Simultaneous Object Localization and Classification |
title_fullStr | A New Loss Function for Simultaneous Object Localization and Classification |
title_full_unstemmed | A New Loss Function for Simultaneous Object Localization and Classification |
title_short | A New Loss Function for Simultaneous Object Localization and Classification |
title_sort | new loss function for simultaneous object localization and classification |
topic | image classification object detection deep learning deep convolutional neural networks computer vision custom training loop |
url | https://www.mdpi.com/2227-7390/11/5/1205 |
work_keys_str_mv | AT andersanchezchica anewlossfunctionforsimultaneousobjectlocalizationandclassification AT benatugartemendiatelleria anewlossfunctionforsimultaneousobjectlocalizationandclassification AT ekaitzzulueta anewlossfunctionforsimultaneousobjectlocalizationandclassification AT unaifernandezgamiz anewlossfunctionforsimultaneousobjectlocalizationandclassification AT javiermariagomezhidalgo anewlossfunctionforsimultaneousobjectlocalizationandclassification AT andersanchezchica newlossfunctionforsimultaneousobjectlocalizationandclassification AT benatugartemendiatelleria newlossfunctionforsimultaneousobjectlocalizationandclassification AT ekaitzzulueta newlossfunctionforsimultaneousobjectlocalizationandclassification AT unaifernandezgamiz newlossfunctionforsimultaneousobjectlocalizationandclassification AT javiermariagomezhidalgo newlossfunctionforsimultaneousobjectlocalizationandclassification |