Multi-Frame Text Aggregation for OCR Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional OCR systems fail to accurately recognize text due to poor lighting, contrast, and focus conditions, and are limited by the need for a minimum character size, which restricts their ability to capture and process large volumes of text in a single image frame, especially in real-world documents.
Innovation Solution
The system aggregates OCR results from multiple image frames over time, using techniques like RANSAC and affine transforms to stitch together text from overlapping regions, improving accuracy and enabling the recognition of larger documents by generating a more comprehensive digital representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single image frame is used for OCR, then the system is simple and fast, but text recognition accuracy deteriorates due to poor lighting, contrast, and focus conditions
Solution Approach 1:
The patent segments the text capture process into multiple image frames captured at different positions, lighting conditions, or focal depths. Each frame captures a portion of the text under different conditions, and the OCR system processes each frame separately before aggregating results. This segmentation allows the system to overcome limitations of single-frame OCR by combining multiple partial observations into a complete and accurate text recognition.
2Quantity of substance
If a single image frame is used, then processing is fast, but the volume of text that can be captured is limited by character size requirements
Solution Approach 1:
The patent transitions from capturing text in a single two-dimensional image frame to capturing text across multiple dimensions by taking several images at different positions, angles, or focal planes. This dimensional expansion allows the system to capture larger volumes of text that would not fit in a single frame while maintaining character size requirements for OCR recognition.
3Reliability
If multiple image frames are aggregated, then text recognition accuracy and volume improve, but system complexity and processing time increase
Solution Approach 1:
The patent performs preliminary actions by capturing multiple image frames in advance under different conditions before the OCR processing begins. The images are pre-positioned, pre-focused, and pre-lit to optimize text visibility. This preliminary capture phase allows the subsequent OCR aggregation to work with already-optimized data, improving overall processing efficiency despite handling multiple frames.
Data Source
AI summary
Various embodiments enable text aggregation from multiple image frames of text. Accordingly, in order to stitch newly scanned areas of a document together, text in a respective image is recognized and analyzed using an algorithm to identify pairs of corresponding words in other images. Upon identifying a minimum number of matching pairs between two respective images, a mapping between the same can be determined based at least in part on a geometric correspondence between respective identified pairs. Based on this mapping, the recognized text of the two images can be merged by adding words of one image to the other using the matching word pairs as alignment data points.