Vehicle Brand Identification Using Text Accumulation and Logo Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for identifying vehicle brands and models require frequent retraining when new vehicles are introduced or existing vehicles modify their shapes, logos, or license plates, and consume high computational resources with high latency.
Innovation Solution
A computing system that uses a processor and memory to obtain video frames from video capturing devices, detecting text or logos using convolution neural networks to identify vehicle brands and models, employing image processing and character recognition techniques to accumulate and match recognized text with a mapping table.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing techniques use logo-based identification with convolution neural networks, then brand identification accuracy is improved, but computational resources and processing time increase significantly
Solution Approach 1:
The system segments the identification task into two distinct paths: text-based OCR identification and logo-based CNN identification. This segmentation allows the system to choose the more efficient text-based approach when applicable, reducing computational resource usage while maintaining high identification accuracy.
Solution Approach 2:
The system changes the identification parameter from purely visual logo recognition to text-based character recognition when text is present. This parameter change from image-based CNN processing to text-based OCR processing significantly reduces computational complexity and energy consumption while maintaining or improving identification accuracy.
2Measurement precision
If existing techniques retrain models for new vehicles or modified shapes, then identification accuracy is maintained, but time and computational overhead increase
Solution Approach 1:
The text-based OCR system is inherently adaptable to new vehicles and modifications without requiring retraining. The system automatically adjusts to new brand names, model names, and text formats through its existing character recognition capabilities, eliminating the need for time-consuming retraining processes while maintaining high identification accuracy.
3Adaptability or versatility
If existing techniques process multiple vehicle features, then comprehensive identification is achieved, but system complexity increases
Solution Approach 1:
The system dynamically switches between text-based and logo-based identification methods based on the presence of text in the captured image. This dynamic approach simplifies the system architecture by providing a clear decision pathway while maintaining comprehensive identification coverage through the dual-method capability.
Data Source
AI summary
The present invention relates to a method of identifying a brand and a model of a vehicle. The method comprises obtaining one or more video frames. Further, the method comprises detecting one of presence or absence of text in the at least one vehicle. Upon detecting the presence of the text, the method comprises determining at least one of the brand and the model of the at least one vehicle by accumulating the text recognized in the one or more video frames. Further, identifying at least one of the brand and the model of the at least one vehicle based on an accumulated text. Upon detecting the absence of the text, the method comprises determining the brand of the at least one vehicle based on logo associated with the vehicle.