Image Recognition System Integrating Knowledge Bases for Text Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optical Character Recognition (OCR) technologies face challenges in accurately recognizing text from images, especially in complex scenes, as they often rely solely on image information without incorporating relevant knowledge information, leading to reduced accuracy in applications like medicine name recognition.
Innovation Solution
An image recognition method that combines feature extraction from images using convolution neural networks with knowledge representation data from knowledge bases, employing recurrent neural networks to fuse image and knowledge information, improving text recognition accuracy by determining relevant knowledge data based on similarity parameters and concatenating features to predict text content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OCR technology relies solely on image information for text recognition, then the system complexity is low, but the text recognition accuracy deteriorates in complex scenes
Solution Approach 1:
The patent merges image information from the OCR system with knowledge information from an external knowledge base. The text recognition process combines visual features extracted from images with semantic knowledge about text content (such as medicine name information), creating a fused representation that improves recognition accuracy in complex scenes by leveraging both visual and knowledge-based cues.
2Measurement precision
If knowledge information is integrated into the text recognition process, then text recognition accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent introduces an embedding layer as an intermediary component that bridges the image processing pipeline and the knowledge base. This embedding layer transforms text into vector representations and facilitates the integration of knowledge information without requiring fundamental changes to the existing OCR architecture, thereby managing system complexity while improving accuracy.
3Measurement precision
If only image information is used for text recognition, then the processing speed is fast, but the recognition accuracy in complex scenes deteriorates
Solution Approach 1:
The patent performs preliminary actions by pre-processing and embedding knowledge information into vector representations before the actual text recognition process. The knowledge base is prepared in advance with structured data about text content (such as medicine names, specifications), which are then efficiently retrieved and integrated during recognition, reducing the computational burden during real-time processing while maintaining high accuracy.
Data Source
AI summary
An image recognition method and apparatus, a training method, an electronic device, and a storage medium are provided. The image recognition method includes: acquiring an image to be recognized, the image to be recognized including a target text; and determining text content of the target text based on knowledge information and image information of the image to be recognized.