Logo Picture Recognition Using Text and Graph Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing logo picture recognition technologies suffer from low accuracy due to the exclusion of text information during matching, leading to incorrect identification of similar logo graphs with different text.
Innovation Solution
A method that includes obtaining a logo picture with both graph and text information, performing text recognition, and searching for matches based on both the logo graph and text, using techniques like OCR and deep neural networks to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If picture matching is performed only according to logo graph features, then the matching process is simple and fast, but the recognition accuracy is low due to incorrect matches between similar graphs with different text
Solution Approach 1:
The patent segments the logo picture into two independent components: logo graph features and text features. The matching process is divided into separate stages - first matching logo graphs, then verifying text information. This segmentation allows the system to incorporate multiple feature types without creating a single complex matching algorithm, thereby improving accuracy while managing complexity through modular processing
Solution Approach 2:
The patent introduces text information as an intermediary verification layer between logo graph matching and final recognition results. After initial logo graph matching, text recognition results serve as a mediator to confirm or reject potential matches. This intermediary step resolves the contradiction by adding accuracy without requiring complete redesign of the matching process
2Productivity
If text information is excluded during matching, then the matching process is faster and simpler, but incorrect identification occurs between similar logo graphs with different text
Solution Approach 1:
The patent performs text recognition as a preliminary action alongside logo graph extraction, before the matching process begins. Both text and graph features are prepared in advance, allowing the matching phase to proceed efficiently by comparing pre-processed features rather than processing raw images during matching. This preliminary preparation maintains speed while enabling accurate text-based verification
Solution Approach 2:
The patent applies partial text verification rather than requiring complete text matching for all cases. The system performs text recognition and comparison, but uses it as one of multiple verification criteria rather than an absolute requirement. This partial application of text checking maintains high productivity while sufficiently improving reliability by filtering out obvious mismatches
Data Source
AI summary
The present disclosure provides a logo picture processing method, apparatus, device and medium, and relates to technical field of image processing, and specifically to the technical field of artificial intelligence such as deep learning and computer vision. The logo picture processing method includes: obtaining a logo picture including: a current logo graph and current text information; performing text recognition on the logo picture to obtain the current text information; searching for a picture that matches both the current logo graph and the current text information, to obtain a matched picture. The present disclosure may improve the accuracy of the matched picture of the logo picture and thereby improve the logo picture recognition accuracy.