Automatic license plate recognition with convolutional neural networks trained on synthetic data

Björklund, Tomas; Fiandrotti, Attilio; Annarumma, Mauro; Francini, Gianluca; Magli, Enrico

doi:10.1109/mmsp.2017.8122260

Cited by 22 publications

(10 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the last step, a bidirectional recurrent neural network (BRNN) with Connectionist Temporal Classification recognizes the LP characters. Björklund [33] proposed an ALPR system trained on synthetic data that has varying pose conditions and illumination levels and showed precision and recall of 93%. Table 1 presents the overall literature review of LP detection, LP region of interest extraction, characters’ segmentation, and character recognition.…”

Section: Related Workmentioning

confidence: 99%

“…zoningfeature=∑i=1M∑j=1Nboldsubimage(i,j) where M×N is the size of sub-image. We considered a character 42×24 size of image and then divided it into nine sub-images of 14×8 each. Perimeter: The set of interior boundary pixels of a connected component (character image ( C )) [33]. We considered 8-connectivity to find the perimeter.…”

Section: Proposed Systemmentioning

confidence: 99%

See 1 more Smart Citation

Barrier Access Control Using Sensors Platform and Vehicle License Plate Characters Recognition

Ullah

Anwar

Shahzadi

et al. 2019

Sensors

View full text Add to dashboard Cite

The paper proposes a sensors platform to control a barrier that is installed for vehicles entrance. This platform is automatized by image-based license plate recognition of the vehicle. However, in situations where standardized license plates are not used, such image-based recognition becomes non-trivial and challenging due to the variations in license plate background, fonts and deformations. The proposed method first detects the approaching vehicle via ultrasonic sensors and, at the same time, captures its image via a camera installed along with the barrier. From this image, the license plate is automatically extracted and further processed to segment the license plate characters. Finally, these characters are recognized with the help of a standard optical character recognition (OCR) pipeline. The evaluation of the proposed system shows an accuracy of 98% for license plates extraction, 96% for character segmentation and 93% for character recognition.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Proposed Systemmentioning

confidence: 99%

Barrier Access Control Using Sensors Platform and Vehicle License Plate Characters Recognition

Ullah

Anwar

Shahzadi

et al. 2019

Sensors

View full text Add to dashboard Cite

show abstract

“…The problem of detecting the expiry dates from images shares many similarities with the problem of reading car licence plate from traffic cameras. One interesting approach to this problem is shown in [8] where the authors train a neural network on an artificially enlarged dataset but that work differs greatly from our approach as we take a general purpose OCR software and study the impact of image pre-processing algorithms on its performance. Usage of Hough transform to generally improve OCR performance outside of the specific problem of extracting expiry dates has already been implemented in [9], [10].…”

Section: Related Workmentioning

confidence: 99%

Usage of Hough Transform for Expiry Date Extraction via Optical Character Recognition

Scazzoli

Bartezzaghi

Uysal

et al. 2019

2019 Advances in Science and Engineering Technology International Conferences (ASET)

View full text Add to dashboard Cite

In this article we evaluate the impact of using two image pre-processing approaches with the objective of aiding an Optical Character Recognition (OCR) software in correctly retrieving an expiry date from an image of a product containing it. In particular, we analyze the impact of finding the rotation angle of an image using the Hough transform and the impact of image binarization using adaptive Gaussian threshold. We attempt to further increase OCR accuracy through a sliding window approach. Our results show that applying the Hough transform noticeably improves OCR performance with minimal impact on the execution time.

show abstract

“…In this paper, a simplified flow for Italian license plate recognition on an embedded platform is presented. A neural network-based embedded framework is used to replicate a highly accurate automatic license plate recognition (ALPR) system [13,14]. Trained parameters are imported and used by a CUDA-based replicated network.…”

Section: Related Workmentioning

confidence: 99%

“…A fully convolutional neural network architecture is trained using a synthetic dataset containing images of a wide variation. It was tested on a dataset of real images to ensure that the learned features are robust to different image capturing conditions [14]. Torch computing framework was used to construct and train this architecture for a license plate recognition system.…”

Section: Network Developed For Desktop and Server Environmentsmentioning

confidence: 99%

Deep Classifiers-Based License Plate Detection, Localization and Recognition on GPU-Powered Mobile Platform

et al. 2017

Self Cite

View full text Add to dashboard Cite

The realization of a deep neural architecture on a mobile platform is challenging, but can open up a number of possibilities for visual analysis applications. A neural network can be realized on a mobile platform by exploiting the computational power of the embedded GPU and simplifying the flow of a neural architecture trained on the desktop workstation or a GPU server. This paper presents an embedded platform-based Italian license plate detection and recognition system using deep neural classifiers. In this work, trained parameters of a highly precise automatic license plate recognition (ALPR) system are imported and used to replicate the same neural classifiers on a Nvidia Shield K1 tablet. A CUDA-based framework is used to realize these neural networks. The flow of the trained architecture is simplified to perform the license plate recognition in real-time. Results show that the tasks of plate and character detection and localization can be performed in real-time on a mobile platform by simplifying the flow of the trained architecture. However, the accuracy of the simplified architecture would be decreased accordingly.

show abstract

Automatic license plate recognition with convolutional neural networks trained on synthetic data

Cited by 22 publications

References 4 publications

Barrier Access Control Using Sensors Platform and Vehicle License Plate Characters Recognition

Barrier Access Control Using Sensors Platform and Vehicle License Plate Characters Recognition

Usage of Hough Transform for Expiry Date Extraction via Optical Character Recognition

Deep Classifiers-Based License Plate Detection, Localization and Recognition on GPU-Powered Mobile Platform

Contact Info

Product

Resources

About