Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture
In recent years, Vehicle Make and Model Recognition (VMMR) has attracted a lot of attention as it plays a crucial role in Intelligent Transportation Systems (ITS). Accurate and efficient VMMR systems are required in real-world applications including intelligent surveillance and autonomous driving. T...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-11-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/21/8439 |
_version_ | 1797466455611015168 |
---|---|
author | Yangxintong Lyu Ionut Schiopu Bruno Cornelis Adrian Munteanu |
author_facet | Yangxintong Lyu Ionut Schiopu Bruno Cornelis Adrian Munteanu |
author_sort | Yangxintong Lyu |
collection | DOAJ |
description | In recent years, Vehicle Make and Model Recognition (VMMR) has attracted a lot of attention as it plays a crucial role in Intelligent Transportation Systems (ITS). Accurate and efficient VMMR systems are required in real-world applications including intelligent surveillance and autonomous driving. The paper introduces a new large-scale dataset and a novel deep learning paradigm for VMMR. A new large-scale dataset dubbed Diverse large-scale VMM (DVMM) is proposed collecting image-samples with the most popular vehicle brands operating in Europe. A novel VMMR framework is proposed which follows a two-branch architecture performing make and model recognition respectively. A two-stage training procedure and a novel decision module are proposed to process the make and model predictions and compute the final model prediction. In addition, a novel metric based on the true positive rate is proposed to compare classification confusion of the proposed 2B–2S and the baseline methods. A complex experimental validation is carried out, demonstrating the generality, diversity, and practicality of the proposed DVMM dataset. The experimental results show that the proposed framework provides <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>93.95</mn><mo>%</mo></mrow></semantics></math></inline-formula> accuracy over the more diverse DVMM dataset and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>95.85</mn><mo>%</mo></mrow></semantics></math></inline-formula> accuracy over traditional VMMR datasets. The proposed two-branch approach outperforms the conventional one-branch approach for VMMR over small-, medium-, and large-scale datasets by providing lower vehicle model confusion and reduced inter-make ambiguity. The paper demonstrates the advantages of the proposed two-branch VMMR paradigm in terms of robustness and lower confusion relative to single-branch designs. |
first_indexed | 2024-03-09T18:40:07Z |
format | Article |
id | doaj.art-6c2e9cc647604ce9b6664989eda0b9b9 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-09T18:40:07Z |
publishDate | 2022-11-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-6c2e9cc647604ce9b6664989eda0b9b92023-11-24T06:48:05ZengMDPI AGSensors1424-82202022-11-012221843910.3390/s22218439Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning ArchitectureYangxintong Lyu0Ionut Schiopu1Bruno Cornelis2Adrian Munteanu3Department of Electronics and Informatics, Vrije Universiteit Brussel, 1050 Brussels, BelgiumDepartment of Electronics and Informatics, Vrije Universiteit Brussel, 1050 Brussels, BelgiumDepartment of Electronics and Informatics, Vrije Universiteit Brussel, 1050 Brussels, BelgiumDepartment of Electronics and Informatics, Vrije Universiteit Brussel, 1050 Brussels, BelgiumIn recent years, Vehicle Make and Model Recognition (VMMR) has attracted a lot of attention as it plays a crucial role in Intelligent Transportation Systems (ITS). Accurate and efficient VMMR systems are required in real-world applications including intelligent surveillance and autonomous driving. The paper introduces a new large-scale dataset and a novel deep learning paradigm for VMMR. A new large-scale dataset dubbed Diverse large-scale VMM (DVMM) is proposed collecting image-samples with the most popular vehicle brands operating in Europe. A novel VMMR framework is proposed which follows a two-branch architecture performing make and model recognition respectively. A two-stage training procedure and a novel decision module are proposed to process the make and model predictions and compute the final model prediction. In addition, a novel metric based on the true positive rate is proposed to compare classification confusion of the proposed 2B–2S and the baseline methods. A complex experimental validation is carried out, demonstrating the generality, diversity, and practicality of the proposed DVMM dataset. The experimental results show that the proposed framework provides <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>93.95</mn><mo>%</mo></mrow></semantics></math></inline-formula> accuracy over the more diverse DVMM dataset and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>95.85</mn><mo>%</mo></mrow></semantics></math></inline-formula> accuracy over traditional VMMR datasets. The proposed two-branch approach outperforms the conventional one-branch approach for VMMR over small-, medium-, and large-scale datasets by providing lower vehicle model confusion and reduced inter-make ambiguity. The paper demonstrates the advantages of the proposed two-branch VMMR paradigm in terms of robustness and lower confusion relative to single-branch designs.https://www.mdpi.com/1424-8220/22/21/8439deep-learningVehicle Make and Model Recognitiontwo-branch strategylarge-scale vehicle datasetIntelligent Transportation System (ITS) |
spellingShingle | Yangxintong Lyu Ionut Schiopu Bruno Cornelis Adrian Munteanu Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture Sensors deep-learning Vehicle Make and Model Recognition two-branch strategy large-scale vehicle dataset Intelligent Transportation System (ITS) |
title | Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture |
title_full | Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture |
title_fullStr | Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture |
title_full_unstemmed | Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture |
title_short | Framework for Vehicle Make and Model Recognition—A New Large-Scale Dataset and an Efficient Two-Branch–Two-Stage Deep Learning Architecture |
title_sort | framework for vehicle make and model recognition a new large scale dataset and an efficient two branch two stage deep learning architecture |
topic | deep-learning Vehicle Make and Model Recognition two-branch strategy large-scale vehicle dataset Intelligent Transportation System (ITS) |
url | https://www.mdpi.com/1424-8220/22/21/8439 |
work_keys_str_mv | AT yangxintonglyu frameworkforvehiclemakeandmodelrecognitionanewlargescaledatasetandanefficienttwobranchtwostagedeeplearningarchitecture AT ionutschiopu frameworkforvehiclemakeandmodelrecognitionanewlargescaledatasetandanefficienttwobranchtwostagedeeplearningarchitecture AT brunocornelis frameworkforvehiclemakeandmodelrecognitionanewlargescaledatasetandanefficienttwobranchtwostagedeeplearningarchitecture AT adrianmunteanu frameworkforvehiclemakeandmodelrecognitionanewlargescaledatasetandanefficienttwobranchtwostagedeeplearningarchitecture |