OCR Anchor Identification for Document Field Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in efficiently identifying and extracting data from scanned documents due to issues like optical character recognition errors, text differences, graphics, noise, rotations, skewing, and handwriting, which degrade image quality and complicate the identification of anchors for fields in documents.
Innovation Solution
A method using optical character recognition data to identify anchors for fields by collecting characters associated with specific words, creating anchors based on these collections, and applying them to identify corresponding fields in other documents, thereby enhancing the probability of matching and extracting data accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional OCR methods are used to extract data from scanned documents, then the process is simple and fast, but the accuracy deteriorates due to OCR errors, text differences, graphics, noise, rotations, skewing, and handwriting
Solution Approach 1:
The patent segments the document processing into distinct phases: OCR data collection, anchor identification, and field extraction. By dividing the complex task of accurate data extraction from degraded documents into manageable segments, the system can apply specialized techniques at each stage without overwhelming complexity
Solution Approach 2:
The patent introduces anchors as intermediary elements that mediate between the degraded OCR data and the target fields. These anchors serve as reference points that bridge the gap between noisy OCR output and accurate field identification, enabling precise data extraction even when direct OCR matching fails
2Reliability
If multiple characters and words are used to create anchors, then the robustness of anchor identification improves, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary OCR data collection and anchor identification before the actual field extraction process. By pre-processing the document to identify reliable anchors using multiple characters and words, the system establishes a robust foundation that speeds up subsequent data extraction operations
Solution Approach 2:
The patent adjusts parameters such as the number of characters per anchor, search radius, and matching thresholds based on document characteristics. This dynamic parameter adjustment allows the system to maintain high reliability while optimizing processing time for different document types and quality levels
Data Source
AI summary
Identifying anchors for fields using optical character recognition data is described. A collection of characters is identified. The collection of characters includes a first set of characters at a first position relative to a first field in a first document and a second set of characters at a second position relative to the first field in the first document. The first set of characters is associated with a first word, and the second set of characters is associated with a second word. An anchor is created based on the collection of characters, wherein the anchor is at a third relative position to the first field in the first document. A second field is identified in a second document by identifying the anchor in the second document.