Document Skew Correction Interface with Dynamic Candidate Line Display
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in accurately capturing frontal images of documents using mobile terminals due to challenges in maintaining camera alignment and detecting document edges, leading to skewed images that require correction before printing.
Innovation Solution
An information processing apparatus and method that allows users to easily designate quadrilateral areas for skew correction by displaying candidate lines only for the selected side, enabling intuitive selection and correction without cluttering the screen with unnecessary candidates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If all contour candidate sides are displayed on the image when switching among contour candidates, then the user can see all available options, but the displayed image becomes convoluted with extremely low visibility
Solution Approach 1:
The patent divides the contour candidates into multiple groups based on their spatial relationships. Instead of displaying all candidates simultaneously, the system segments them and displays one group at a time, allowing users to navigate through different groups using operation buttons. This segmentation resolves the contradiction by maintaining visibility while preserving access to all candidates through sequential display.
Solution Approach 2:
The system implements periodic display of contour candidate groups, where different groups are displayed in sequence rather than all at once. Users can switch between groups using operation buttons, creating a periodic action pattern that allows comprehensive candidate exploration without overwhelming the display with all candidates simultaneously, thus maintaining image visibility.
2Measurement precision
If the user directly designates the position of vertex and side, then the precision of quadrilateral area designation is improved, but the operation complexity and user effort increase
Solution Approach 1:
The system performs automatic contour detection and generates candidate quadrilateral areas based on edge information from the captured image. This self-service approach allows the system to automatically identify potential document boundaries, reducing the user's workload from manually positioning all four vertices to simply selecting from pre-computed candidates or making minor adjustments, thereby maintaining precision while reducing operation effort.
Solution Approach 2:
The system performs preliminary contour detection and candidate generation before user selection. By pre-processing the image to identify edge information and compute multiple candidate quadrilateral areas, the system prepares the groundwork for accurate designation, allowing users to make informed selections without performing complex manual positioning tasks, thus achieving precision with reduced effort.
3Ease of operation
If automatic detection of document area sides using edge information is used, then the processing load on the user is minimized, but the detection may fail when edges cannot be detected due to low contrast or excessive edges
Solution Approach 1:
The system introduces an intermediary manual selection mechanism between automatic detection and final document area designation. When automatic edge detection fails or produces unreliable results due to low contrast or excessive edges, users can intervene to manually select from candidate areas or adjust the detected contours. This intermediary approach maintains low processing load while compensating for detection failures, ensuring reliability.
Solution Approach 2:
The system implements a dynamic approach where the detection method can switch between automatic and manual modes based on detection confidence. When edge detection succeeds with high confidence, the system operates automatically with minimal user involvement. When detection fails or confidence is low, the system dynamically transitions to allow manual intervention, thus maintaining low processing load overall while ensuring reliability when needed.
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
A detected quadrilateral area is displayed and no group of candidate lines is displayed in a normal state. While a user is selecting a side that the user desires to change, a group of candidate lines corresponding to the selected side is displayed. Then, whether to replace a position of the selected side with a position of a candidate line is determined based on a movement destination position of the selected side.