Structured Multi-Phase Image Review With Text and Graphics Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current review processes for complex images, such as CAD drawings with overlaid text and graphics, are unstructured, leading to increased errors due to the need for mental tracking of reviewed areas and difficulties in viewing detailed information, which results in lower confidence and higher error rates.
Innovation Solution
A structured, multi-phase review process that analyzes images to detect and sequentially display portions containing text and graphics, optionally rotating and enlarging them for easier review, allowing users to provide feedback and make revisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reviewers manually track reviewed areas in complex images, then they can review the entire image, but error rates increase and confidence decreases
Solution Approach 1:
The system automatically segments the complex image into distinct regions (e.g., text regions, graphic regions, CAD drawing regions) and guides the reviewer through each region sequentially. This eliminates the need for mental tracking by structuring the review process into manageable segments, thereby reducing errors and increasing confidence.
Solution Approach 2:
The system provides structured feedback to the reviewer by indicating which regions have been reviewed and which remain. This feedback mechanism replaces manual mental tracking with an automated system that tracks progress and guides the reviewer through the review process, improving reliability.
2Loss of information
If detailed information in complex images is viewed at original scale, then complete information is visible, but viewing difficulty increases
Solution Approach 1:
The system extracts text and graphics from the complex image and displays them in a separate dimensional space (overlaying them on the image or displaying them in dedicated panels). This allows detailed information to be viewed at enlarged scales without losing context, as the overlay or separate display maintains the relationship between detailed elements and the overall image.
3Productivity
If unstructured review process is used, then review speed is maintained, but error rates increase
Solution Approach 1:
The system performs preliminary actions by automatically detecting and organizing regions of interest before the reviewer begins. Text regions, graphic regions, and other elements are pre-identified and structured, so the reviewer can proceed through a systematic sequence without losing speed. This preliminary structuring prevents errors while maintaining productivity.
4Loss of information
If text and graphics are overlaid on CAD drawings, then complete information is presented, but detection and measurement difficulty increases
Solution Approach 1:
The system automatically segments the overlaid elements (text, graphics, CAD drawings) into distinct detectable regions. By identifying and separating these elements, the system makes each detectable and measurable independently, eliminating the difficulty caused by overlapping while preserving all information.
Solution Approach 2:
The system introduces an intermediary processing layer that analyzes the complex overlaid image and extracts individual elements. This intermediary step detects text regions, graphic regions, and CAD elements separately, making them measurable and reviewable without the confusion of direct overlay inspection.
Data Source
AI summary
The devices, systems, and methods described herein are directed to obtaining user feedback regarding an image containing text and graphics. In some examples, the image is analyzed to detect one or more portions of the image that contain text, which are sequentially displayed to a user. The user provides feedback regarding the text. Each of the portions of the image that contain graphics are displayed to the user, and the user provides feedback regarding the graphics. In some examples, the text and/or graphics are rotated and/or enlarged before being displayed to the user.