PlateSegFL: A Privacy-Preserving License Plate Detection Using Federated Segmentation Learning

Md. Shahriar Rahman Anuvab,Mishkat Sultana,Md. Atif Hossain,Shashwata Das,Suvarthi Chowdhury,Rafeed Rahman,Dibyo Fabian Dofadar,Shahriar Rahman Rana

2024-04-08

Abstract:Automatic License Plate Recognition (ALPR) is an integral component of an intelligent transport system with extensive applications in secure transportation, vehicle-to-vehicle communication, stolen vehicles detection, traffic violations, and traffic flow management. The existing license plate detection system focuses on one-shot learners or pre-trained models that operate with a geometric bounding box, limiting the model's performance. Furthermore, continuous video data streams uploaded to the central server result in network and complexity issues. To combat this, PlateSegFL was introduced, which implements U-Net-based segmentation along with Federated Learning (FL). U-Net is well-suited for multi-class image segmentation tasks because it can analyze a large number of classes and generate a pixel-level segmentation map for each class. Federated Learning is used to reduce the quantity of data required while safeguarding the user's privacy. Different computing platforms, such as mobile phones, are able to collaborate on the development of a standard prediction model where it makes efficient use of one's time; incorporates more diverse data; delivers projections in real-time; and requires no physical effort from the user; resulting around 95% F1 score.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper discusses how to use Federated Learning and U-Net segmentation technique to achieve privacy-preserving license plate detection. Current license plate recognition systems typically rely on one-time learning models or pre-trained models, which use geometric bounding boxes and limit the model's performance. In addition, continuously uploading video data streams to a central server leads to network and complexity issues.

To address these problems, the paper proposes a method called PlateSegFL, which combines U-Net (suitable for multi-class image segmentation tasks) and Federated Learning. Federated Learning can reduce data requirements while protecting user privacy, allowing different computing platforms (such as mobile devices) to collaborate in building standard prediction models, improving efficiency, achieving real-time predictions, and reducing user involvement. This method conducts model training without compromising privacy, especially for handling irregular mask problems, and achieves an F1 score of approximately 95% in license plate recognition.

The paper also compares different license plate detection methods such as the YOLO series and U-Net, pointing out the limitations of YOLO in handling irregular masks and improving these issues by introducing Federated Learning and U-Net. The study used multiple datasets, including license plate data from Europe, the Americas, Tunisia, and Bangladesh, to train and test the model. The data set size was expanded using data augmentation, and a series of statistical metrics were used to evaluate model performance.

Overall, this paper aims to create an efficient and privacy-preserving license plate recognition system by combining Federated Learning and U-Net, especially suitable for real-time monitoring and traffic management to address the increasing problem of vehicle-related crimes.