Document Image Capture via Block Segmentation and Parameter Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current optical image capture systems face challenges in capturing high-quality images of large documents, leading to inefficient data extraction due to suboptimal focus and exposure settings across the document, requiring multiple image captures and increased bandwidth and processing power, while also posing security risks.

Innovation Solution

A system and method that divides the document into image blocks, optimizes image capture parameters like focus and exposure for each block, captures and aggregates these images to reconstruct a high-quality document image, and transfers it efficiently to a data extraction engine.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If the camera is positioned a significant distance away from the source document to capture the entire document, then the entire document can be captured in the image, but the image capture parameters (focus and exposure) are only optimized for a portion of the document, resulting in poor quality images of other portions

Engineering Contradiction:
Improvedocument coverage areaVSAvoidimage quality
Core Design Contradiction:
Area of stationary objectVSManufacturing precision

Solution Approach 1:

The patent divides the document into multiple image blocks (e.g., quadrants or regions) and captures each block separately with optimized camera parameters for that specific region. This segmentation allows the camera to maintain high focus and exposure settings for each individual block while still covering the entire document across multiple captures.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different image capture parameters (focus, exposure, white balance) to different regions of the document based on local lighting conditions and document characteristics. Each image block is captured with parameters optimized for its specific location, ensuring high image quality across the entire document surface.

Inventive Principle:
Principle #3Local quality

2Reliability

If multiple images of the source document are captured to ensure all portions are clearly captured, then complete document data can be obtained, but the bandwidth required to transfer multiple images and the processor cycles required to process them increase significantly

Engineering Contradiction:
Improvedata extraction accuracyVSAvoidbandwidth and processing power
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The patent extracts only the necessary image blocks containing relevant data from the document for processing and transfer. Instead of transferring complete multiple images, the system identifies and extracts specific regions of interest (image blocks) that contain the required information, reducing the total data volume that needs to be transmitted and processed.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary processing of captured images to identify and extract only the necessary image blocks containing relevant data before transfer. This preliminary action reduces the amount of data that needs to be transmitted and processed by the data extraction engine, conserving bandwidth and processing power.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If multiple images of the source document are transferred to a remote data extraction engine, then complete document data can be processed, but security risks increase particularly for sensitive documents

Engineering Contradiction:
Improvedata extraction completenessVSAvoidsecurity risk
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent extracts and processes only the necessary image blocks locally or with minimal transfer to remote systems. By extracting only the essential data-containing regions rather than transferring complete multiple images to remote servers, the system reduces the security exposure for sensitive documents while maintaining data extraction completeness.

Inventive Principle:
Principle #2Taking out (Extraction)

4Reliability

If the user is forced to capture multiple images of the source document to obtain complete data, then all document information can be captured, but the process becomes frustrating and time-consuming for the user

Engineering Contradiction:
Improvedata extraction completenessVSAvoiduser effort
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent implements automatic image block capture and processing without requiring manual user intervention. The system automatically divides the document into image blocks, captures each block with optimized parameters, and processes them for data extraction. This self-service approach eliminates the need for users to manually capture multiple images, significantly reducing user effort and frustration.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent performs preliminary automatic processing to determine which image blocks contain relevant data and captures only those necessary blocks. This preliminary action reduces the number of images the user needs to capture and provides a more streamlined, user-friendly experience while ensuring complete data extraction.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2979225B1Method and system for creating optimized images for data identification and extraction
Publication Date: 2018.08.29 INTUIT INC
  • EP2979225B1 patent drawingFigure 1
  • EP2979225B1 patent drawingFigure 2A
  • EP2979225B1 patent drawingFigure 2B

AI summary

A viewfinder screen display is generated and positioned such that a source document is displayed in the viewfinder screen display. Source document image blocks corresponding to different portions of the source document are then defined. For each source document image block, the image capture parameter of an image capture device is set to an optimized image captare parameter setting for the source document image block. The image capture device then captures an image block optimized image of the source document optimized for the source document image block. The optimized source document image blocks are then extracted from each image block optimized image of the source document. The extracted optimized source document image blocks are then aggregated and used to construct an image capture parameter optimized, image of the source document.