Picture Search OCR Caching for Faster Text Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing picture search methods based on optical character recognition (OCR) face inefficiencies due to high time costs for accurate recognition, inability to find detailed text information, and risks associated with cloud recognition such as failure, time-consuming uploads/downloads, and privacy concerns.
Innovation Solution
A picture search method combining low-dimensional and high-dimensional OCR processing to optimize recognition efficiency and accuracy, where low-dimensional OCR is performed based on an OCR threshold and high-dimensional OCR is conducted during idle times, with results stored for subsequent matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complete OCR processing is performed on all pictures to ensure accurate text recognition, then text recognition accuracy is improved, but search time increases significantly
Solution Approach 1:
The patent segments the picture library into two parts: pictures that have pre-stored OCR results and pictures that require OCR processing. This segmentation allows the system to quickly retrieve results for some pictures while processing only the necessary subset, thereby reducing overall search time while maintaining recognition accuracy.
Solution Approach 2:
The patent performs OCR processing in advance and stores the results in the picture library. When a search is conducted, the system can directly retrieve pre-processed OCR results without performing real-time OCR on all pictures, significantly reducing search time while maintaining high text recognition accuracy.
2Productivity
If OCR results are stored in the picture library for faster retrieval, then search efficiency is improved, but storage space consumption increases
Solution Approach 1:
The patent applies local quality by selectively storing OCR results based on picture characteristics and usage patterns. Not all pictures are processed and stored with full OCR results - only those that benefit from it. This selective approach improves search efficiency for relevant pictures while minimizing unnecessary storage consumption.
3Speed
If low-dimensional OCR processing is used to reduce processing time, then search speed is improved, but text recognition accuracy decreases
Solution Approach 1:
The patent implements a dynamic OCR processing strategy where the processing dimension is adjusted based on the specific picture and search requirements. For some pictures, low-dimensional OCR is used for quick retrieval, while for others requiring higher accuracy, full-dimensional OCR is performed. This dynamic adaptation allows the system to optimize between speed and accuracy based on actual needs.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
Disclosed is a picture search method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product, relating to the technical field of artificial intelligence. The method includes: obtain an OCR result of each picture in a preset picture library in response to a picture search request; traverse pictures which are not subjected to low-dimensional OCR processing and high-dimensional OCR processing in the preset picture library, and perform the low-dimensional OCR processing based on an OCR threshold on each of the traversed pictures to obtain a low-dimensional OCR result of each corresponding picture; determining a target picture matching a key character string in the preset picture library according to at least one of the low-dimensional OCR result and the high-dimensional OCR result of each picture; and determining the target picture as a search result of the picture search request, and displaying the search result.