Medical Device Screen Segmentation for Single-Pass OCR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical device controllers display critical information at varying screen locations due to software version or device type differences, leading to ambiguity in masking relevant portions for optical character recognition (OCR), necessitating multiple OCR processes and increased costs.
Innovation Solution
A server generates a mosaic image by copying source regions from medical device controllers to predefined destination regions, allowing a single OCR process to accurately extract text results by comparing selector text results to predefined text, thereby selecting the correct coordinates based on additional information like software version or device identifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple OCR processes are performed with different masks to handle varying screen layouts, then text extraction accuracy is improved, but processing time and costs increase
Solution Approach 1:
The system performs preliminary actions by detecting screen layout characteristics and identifying the appropriate mask category before executing OCR. The server detects whether the screen displays first or second layout types, selects the corresponding mask in advance, and only then performs OCR processing. This preliminary detection and selection process eliminates the need for multiple OCR attempts, reducing processing time while maintaining accurate text extraction.
2Measurement precision
If multiple OCR processes are performed with different masks to handle varying screen layouts, then text extraction accuracy is improved, but processing costs increase
Solution Approach 1:
The server performs preliminary detection of screen layout characteristics and pre-selects the appropriate mask category before OCR processing. By categorizing screens into first or second layout types and selecting corresponding masks in advance, the system avoids performing multiple OCR processes, thereby reducing processing costs and improving overall processing efficiency while maintaining accurate text extraction.
Solution Approach 2:
The system changes the parameter of mask selection based on detected screen layout characteristics. Instead of using a fixed mask or attempting multiple masks, the server dynamically adjusts the mask selection by detecting layout parameters (such as position of critical information) and selecting the appropriate mask category. This parameter-based adaptation improves processing efficiency while maintaining text extraction accuracy.
3Loss of time
If a single OCR process is used without preliminary masking, then processing time is reduced, but text extraction accuracy deteriorates due to irrelevant text or graphics
Solution Approach 1:
The system performs preliminary masking action before OCR processing by detecting screen layout characteristics and applying the appropriate mask category. This preliminary step removes irrelevant text or graphics that would confuse the OCR engine, ensuring accurate text extraction. The mask is applied in advance based on detected layout parameters, so the subsequent single OCR process operates on a cleaned image, maintaining high accuracy while minimizing processing time.
4Measurement precision
If the server uses several masks and performs OCR with each mask, then text extraction accuracy is improved, but device complexity increases
Solution Approach 1:
The server performs preliminary detection of screen layout characteristics and categorizes the screen into first or second layout types. Based on this categorization, the server selects the appropriate mask category in advance. This preliminary classification simplifies the system architecture by replacing the complex approach of trying multiple masks with multiple OCR processes, with a simpler flow of: detect layout → select mask category → perform single OCR. The complexity is reduced while maintaining text extraction accuracy.
Solution Approach 2:
The system changes the approach from using multiple masks with multiple OCR processes to using parameter-based mask selection. The server detects layout parameters (such as position of critical information) and selects the appropriate mask category based on these parameters. This parameter-driven approach simplifies the system by eliminating the need for multiple OCR processes, reducing device complexity while maintaining accurate text extraction.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A medical device monitoring system and method extract information from screen images from medical device controllers, with a single OCR process invocation per screen image, despite critical information appearing in different screen locations, depending on which medical device controller's screen image is processed. For example, different software versions of the medical device controllers might display the same type of information in different screen locations. Copies of the critical screen information, one copy from each different screen location, are made in a mosaic image, and then the mosaic image is OCR processed to produce text results. Text is selectively extracted from the OCR text results, depending on contents of a selector field on the screen image, such as a software version number or a heart pump model identifier.