Scribd
Upload
19 views3 pages

Main

The document proposes a deep learning-based license plate recognition system using YOLO v3 object detection and a character recognition model. The system will be trained on license plate images and evaluated on a benchmark dataset compared to traditional computer vision techniques. Future work will improve performance with advanced techniques and more diverse datasets.

Uploaded by

Ashok
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
19 views3 pages

Main

The document proposes a deep learning-based license plate recognition system using YOLO v3 object detection and a character recognition model. The system will be trained on license plate images and evaluated on a benchmark dataset compared to traditional computer vision techniques. Future work will improve performance with advanced techniques and more diverse datasets.

Uploaded by

Ashok
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 3

LicensePlate Recognitionusing YOLO v3 and

Deep Learning
Ashok Kunwar

Abstract
License Plate Recognition (LPR) is a crucial technology for various
applications, such as traÓc monitoring, security and surveillance, and
toll collection. In this proposal, we propose a deep learning-based LPR
system using YOLO v3, a state-of-the-art object detection model. The
system will be trained on a large dataset of license plate images and
will be capable of recognizing license plates in real-time with high ac-
curacy. The proposed system will be compared with existing LPR
systems based on traditional computer vision techniques and the re-
sults will be analyzed to demonstrate the superiority of the proposed
deep learning-based approach.

1 Introduction
License Plate Recognition (LPR) is the process of automatically reading
the license plate number of a vehicle using computer vision techniques. It
is a crucial technology for various applications, such as traÓc monitoring,
security and surveillance, and toll collection. LPR systems are typically
composed of three main components: image acquisition, image processing,
and character recognition. Image acquisition captures an image of the license
plate, image processing pre-processes the image to extract the license plate
region, and character recognition uses optical character recognition (OCR)
techniques to extract the text from the license plate region.
In recent years, deep learning has revolutionized the Ñeld of computer
vision and has shown great promise in various image recognition tasks, in-
cluding object detection. Object detection models, such as YOLO v3, are
trained end-to-end to detect objects in an image and can be used to perform
license plate recognition by detecting the license plate region and recognizing
the characters in the region.

1
2 Proposed Method
The proposed LPR system will be based on YOLO v3, a state-of-the-art
object detection model. YOLO v3 is a deep neural network that is trained
end-to-end to detect objects in an image. The system will be trained on a
large dataset oflicense plate images, which will include a diverse set oflicense
plates with varying backgrounds, illumination conditions, and orientations.
The system will be capable of recognizing license plates in real-time with
high accuracy.
In addition to YOLO v3, the proposed LPR system will also use a char-
acter recognition model, such as a Convolutional Neural Network (CNN),
to extract the text from the license plate region. The character recognition
model will be trained on a large dataset of characters from license plate
images to improve its accuracy in recognizing license plate characters.

3 Experiments and Results


The proposed system will be evaluated on a benchmark dataset of license
plate images. The performance of the system will be compared with existing
LPR systems based on traditional computer vision techniques, such as SVM
and HOG, to demonstrate the superiority of the proposed deep learning-
based approach. The results will be analyzed in terms of accuracy, precision,
recall, and processing time.
In addition to the quantitative analysis, the proposed system will also
be tested in real-world scenarios to evaluate its performance in practical
applications.

4 Conclusion and Future Work


In this proposal, we presented a deep learning-based LPR system using
YOLO v3 and a character recognition model. The proposed system is ex-
pected to outperform existing LPR systems based on traditional computer
vision techniques in terms of accuracy and processing time. The results of
the experiments will provide insights into the potential of deep learning for
license plate recognition and the limitations of the proposed system.
As future work, we plan to further improve the performance of the pro-
posed system by incorporating advanced deep learning techniques, such as
transfer learning and Ñne-tuning, and by using larger and more diverse
datasets. We also plan to extend the proposed system to recognize license
plates in various countries and languages, which will make it more versatile
and applicable to a wider range of applications.

References

2
[1] J. Redmon and A. Farhadi, "YOLOv3: An Incremental Improvement,"
arXiv preprint arXiv:1804.02767 (2018).
[2]S. Jain and S. Bhatia, "A Review on License Plate Recognition Tech-
niques," arXiv preprint arXiv:1712.08585 (2017).
[3]X. Wang and Y. Zhang, "Vehicle license plate recognition using deep
learning techniques," Journal of Ambient Intelligence and Humanized Com-
puting, vol. 8, no. 5, pp. 5023-5032 (2017).

You might also like