MR Workflow Analysis Using Video-Annotated Machine Logs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical imaging systems lack comprehensive information about human activities during imaging procedures, as machine logs and DICOM data do not capture human interactions, leading to incomplete workflow analysis and maintenance insights.
Innovation Solution
Integrate video analysis from cameras positioned to monitor the examination region and imaging bay to detect human activities, extract event annotations, and match these with machine log entries based on timestamps, enabling comprehensive logging of patient and personnel actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If video analysis is integrated to detect human activities during imaging procedures, then information completeness about human interactions is improved, but system complexity and data processing requirements increase
Solution Approach 1:
The patent introduces video analysis technology as an intermediary component that captures human activities during imaging procedures. The video system acts as a mediator between the imaging device and the information system, converting visual observations into structured data that can be integrated with existing machine logs and DICOM data, thereby improving information completeness without directly modifying the core imaging system
Solution Approach 2:
The patent segments the information collection system into distinct components: machine logs for device operations, DICOM data for imaging parameters, and video analysis for human activities. Each component independently captures specific types of information, and the results are integrated through timestamp matching. This segmentation allows the system to improve overall information completeness while managing complexity through modular architecture
2Measurement precision
If extensive video review is performed to analyze human activities, then detection accuracy of activities is improved, but time consumption and processing efficiency worsen
Solution Approach 1:
The patent extracts only the essential information about human activities from the video data through automated analysis algorithms. Instead of requiring reviewers to watch entire video recordings, the system extracts key events such as patient positioning, contrast injection timing, and procedural milestones. This extraction process maintains detection accuracy by focusing on critical moments while dramatically reducing the time required for analysis
Solution Approach 2:
The system performs preliminary video analysis to pre-identify and timestamp important events before they need to be reviewed or referenced. By automatically detecting and marking key activities in advance, the system prepares the data structure so that subsequent analysis or verification can proceed efficiently without requiring time-consuming manual review of entire video sequences
3Loss of information
If video data is stored and reviewed to capture human activities, then information completeness is improved, but data storage requirements and processing load increase
Solution Approach 1:
The patent extracts only the essential event information from video data, converting it into concise text annotations with timestamps. Instead of storing and processing entire video files for analysis, the system extracts key events such as 'patient positioned at T=00:05' or 'contrast injection started at T=00:12'. This extraction reduces data storage requirements from gigabytes of video to kilobytes of structured event data while maintaining information completeness
Solution Approach 2:
The system creates a simplified copy of video information in the form of text-based event annotations with timestamps. Rather than storing the original video data for every analysis task, the system creates lightweight textual representations that capture the essential human activity information. These annotations can be stored and processed efficiently while preserving the key information needed for workflow analysis
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A non-transitory computer readable medium (26) stores instructions readable and executable by at least one electronic processor (20) to perform a method (100) of maintaining a machine log (30) for a medical imaging device (2). The method includes: extracting event annotations (38) from video (17, 18) acquired by at least one video camera (12, 16) positioned to image an examination region of the medical imaging device and/or an imaging bay (3) containing the medical imaging device, the event annotations representing activities occurring during a medical imaging examination performed using the medical imaging device; and recording machine log entries (32) in the machine log of the medical imaging device wherein the machine log entries are generated by the medical imaging device. The recording includes matching the extracted event annotations with machine log entries based on timestamps (40) of the event annotations and timestamps (34) of the machine log entries and annotating the machine log entries with the matched event annotations.