End-to-End License Plate Recognition Model Reducing RAM and Speed Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing license plate recognition technologies face inefficiencies due to redundant computations across multiple deep learning models, leading to slow processing speeds and high RAM usage.
Innovation Solution
An end-to-end lightweight method and apparatus that integrates a pre-trained license plate recognition model comprising a feature extraction network, region candidate localization network, super-resolution generation network, and recurrent neural network, which reuses computational variables to reduce redundancy and improve processing speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple independent deep learning models are used for license plate localization, character segmentation, and character recognition, then each function can be performed with specialized processing, but redundant computations occur and processing speed decreases
Solution Approach 1:
The patent merges multiple independent deep learning models (localization model, segmentation model, recognition model) into a single integrated end-to-end license plate recognition model. This integration allows the model to perform localization, segmentation, and recognition simultaneously without redundant feature extraction, thereby improving processing speed while maintaining recognition accuracy through unified optimization.
Solution Approach 2:
The integrated model serves multiple functions within a single system: it performs license plate localization, character segmentation, and character recognition. By designing a universal model that handles all three tasks, the patent eliminates the need for separate specialized models, reducing computational overhead and improving processing efficiency.
2Reliability
If multiple independent deep learning models are used for license plate recognition, then comprehensive processing is achieved, but a large amount of RAM is consumed due to redundant computational variables
Solution Approach 1:
By combining multiple models into one integrated model, the patent eliminates redundant computational variables that would otherwise be stored in RAM. The unified model shares feature extraction processes across localization, segmentation, and recognition tasks, significantly reducing memory consumption while maintaining comprehensive processing capabilities.
3Reliability
If conventional multi-model license plate recognition is used, then thorough processing is performed, but the system complexity increases due to multiple independent models
Solution Approach 1:
The patent reduces system complexity by merging multiple independent models into a single end-to-end model. This integration simplifies the system architecture, eliminates the need for multiple separate model deployments, and reduces the complexity of model coordination while maintaining thorough processing through the model's multi-functional design.
Data Source
AI summary
Embodiments of the present invention provide an end-to-end lightweight method and apparatus for license plate recognition. The method comprises: obtaining an image to be recognized; obtaining a number of a license plate in the image to be recognized and position coordinates of the license plate in the image to be recognized on the basis of the image to be recognized and a pre-trained target license plate recognition model, wherein the target license plate recognition model comprises a target feature extraction network, a target region candidate localization network, a target super-resolution generation network and a target recurrent neural network. Because in this solution, once an image to be recognized is input into the target license plate recognition model, the target license plate recognition model can output the license plate number and position coordinates of the license plate in the image to be recognized, one realizes an end-to-end model. The model has relatively strong robustness, and it can detect and recognize pictures taken under different camera angles. Moreover, computation variables such as image features can be reused without repeated computations, the model takes up less RAM and the speed of license plate recognition is greatly improved.