Optical Image Quality Feedback via Segmented Test Regions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extracting data from optical images of documents using OCR technology are inefficient due to the need for high-quality images, with users often waiting for entire images to be processed to discover and address quality issues, leading to significant delays and processing inefficiencies.
Innovation Solution
A system and method that divides the optical image of a document into distinct test regions, allowing for initial quality assessment and feedback on image quality without processing the entire image, recommending re-capture if quality is insufficient, and optimizing subsequent scans based on identified issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the entire optical image is processed to ensure complete data extraction, then data extraction completeness is improved, but user wait time and processing efficiency deteriorate
Solution Approach 1:
The patent divides the optical image into multiple test regions (e.g., header, body, footer sections) so that quality assessment can be performed on individual segments rather than the entire image. This allows the system to provide feedback on specific quality issues in particular regions while reducing overall processing time.
Solution Approach 2:
The patent performs preliminary quality assessment on test regions before completing the full OCR processing. By evaluating image quality in advance on selected regions, the system can notify users of potential quality issues before full processing begins, allowing them to address problems proactively.
2Measurement precision
If the entire optical image is scanned and processed, then data extraction accuracy is improved, but processing efficiency and user feedback timeliness deteriorate
Solution Approach 1:
The optical image is segmented into multiple test regions that can be evaluated independently. This segmentation allows the system to assess quality metrics (focus, lighting, distortion) in specific regions without processing the entire image, thereby improving processing efficiency while maintaining accuracy through targeted evaluation.
Solution Approach 2:
The patent implements a feedback mechanism that provides users with real-time information about image quality in specific test regions. This feedback loop allows users to understand quality issues and correct them before full processing, improving overall processing efficiency and data extraction accuracy.
3Power
If remote servers are used for OCR processing, then processing capability is improved, but network dependency and processing delay increase
Solution Approach 1:
The patent performs preliminary quality assessment and test region evaluation locally on the device before sending images to remote servers. This preliminary action reduces the amount of data that needs to be transmitted and processed remotely, thereby reducing network dependency and processing delay while maintaining the benefits of remote processing capability.
Data Source
AI summary
An optical image of a source document is captured. Two or more source document image test regions are then defined/determined. An optical image scan is performed on each source document image test region to determine if there are identifiable alpha-numeric characters or symbols present. If one or more of the source document image test regions are determined not to contain identifiable alpha-numeric characters, the captured optical image of the source document is determined to be of insufficient quality to identify and extract individual characters and symbols and it is recommended that optical images of source documents determined to be of insufficient quality to identify and extract individual characters and symbols be re-captured using an image capture device.