Smart car plate recognition system using multi-task learning

Automatic License Plate Recognition (ALPR) systems are crucial in extracting vehicle information. However, ALPR alone is insufficient for robust vehicle owner identification, especially in the event of misidentification or covered license plates (LPs). Acknowledging the significance of vehicle colou...

Full description

Bibliographic Details
Main Author: Khor, Yin Loon
Format: Final Year Project / Dissertation / Thesis
Published: 2024
Subjects:
Online Access:http://eprints.utar.edu.my/6443/
http://eprints.utar.edu.my/6443/1/3E_1903505_FYP_report_%2D_YIN_LOON_KHOR.pdf
_version_ 1848886680732303360
author Khor, Yin Loon
author_facet Khor, Yin Loon
author_sort Khor, Yin Loon
building UTAR Institutional Repository
collection Online Access
description Automatic License Plate Recognition (ALPR) systems are crucial in extracting vehicle information. However, ALPR alone is insufficient for robust vehicle owner identification, especially in the event of misidentification or covered license plates (LPs). Acknowledging the significance of vehicle colour in enhancing identification accuracy, this project proposes a more secure and comprehensive approach by integrating Vehicle Colour Recognition (VCR) with LP detection and Optical Character Recognition (OCR) tasks. Unlike the conventional two-stage ALPR systems, this solution introduces a novel one�stage YOLO-based multi-task model. It incorporates additional object detection heads onto the YOLO backbone, allowing for parallel processing and efficient real-time detection for all three tasks. The proposed model achieves spectacular results with mean Average Precision (mAP) scores of 0.778, 0.963, and 0.881 for OCR, LP detection, and VCR, respectively. Promisingly, this model is comparable to single-head, single-task models, which are trained solely for each task. It outperforms a single-head multi-task model, which naively shares all tasks using one single head. Specifically, the model is 1.77x faster than the conventional approach, which involves inference of single-task models for OCR, LP, and VCR sequentially. Experimental results demonstrate that the proposed solution is robust in simultaneously addressing OCR, LP detection, and VCR within a unified, single-stage framework.
first_indexed 2025-11-15T19:42:21Z
format Final Year Project / Dissertation / Thesis
id utar-6443
institution Universiti Tunku Abdul Rahman
institution_category Local University
last_indexed 2025-11-15T19:42:21Z
publishDate 2024
recordtype eprints
repository_type Digital Repository
spelling utar-64432024-06-20T10:47:47Z Smart car plate recognition system using multi-task learning Khor, Yin Loon TK Electrical engineering. Electronics Nuclear engineering TL Motor vehicles. Aeronautics. Astronautics Automatic License Plate Recognition (ALPR) systems are crucial in extracting vehicle information. However, ALPR alone is insufficient for robust vehicle owner identification, especially in the event of misidentification or covered license plates (LPs). Acknowledging the significance of vehicle colour in enhancing identification accuracy, this project proposes a more secure and comprehensive approach by integrating Vehicle Colour Recognition (VCR) with LP detection and Optical Character Recognition (OCR) tasks. Unlike the conventional two-stage ALPR systems, this solution introduces a novel one�stage YOLO-based multi-task model. It incorporates additional object detection heads onto the YOLO backbone, allowing for parallel processing and efficient real-time detection for all three tasks. The proposed model achieves spectacular results with mean Average Precision (mAP) scores of 0.778, 0.963, and 0.881 for OCR, LP detection, and VCR, respectively. Promisingly, this model is comparable to single-head, single-task models, which are trained solely for each task. It outperforms a single-head multi-task model, which naively shares all tasks using one single head. Specifically, the model is 1.77x faster than the conventional approach, which involves inference of single-task models for OCR, LP, and VCR sequentially. Experimental results demonstrate that the proposed solution is robust in simultaneously addressing OCR, LP detection, and VCR within a unified, single-stage framework. 2024 Final Year Project / Dissertation / Thesis NonPeerReviewed application/pdf http://eprints.utar.edu.my/6443/1/3E_1903505_FYP_report_%2D_YIN_LOON_KHOR.pdf Khor, Yin Loon (2024) Smart car plate recognition system using multi-task learning. Final Year Project, UTAR. http://eprints.utar.edu.my/6443/
spellingShingle TK Electrical engineering. Electronics Nuclear engineering
TL Motor vehicles. Aeronautics. Astronautics
Khor, Yin Loon
Smart car plate recognition system using multi-task learning
title Smart car plate recognition system using multi-task learning
title_full Smart car plate recognition system using multi-task learning
title_fullStr Smart car plate recognition system using multi-task learning
title_full_unstemmed Smart car plate recognition system using multi-task learning
title_short Smart car plate recognition system using multi-task learning
title_sort smart car plate recognition system using multi-task learning
topic TK Electrical engineering. Electronics Nuclear engineering
TL Motor vehicles. Aeronautics. Astronautics
url http://eprints.utar.edu.my/6443/
http://eprints.utar.edu.my/6443/1/3E_1903505_FYP_report_%2D_YIN_LOON_KHOR.pdf