Medical Image Information Extraction for Privacy and Device Compatibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical information management systems, such as PACS, face challenges with patient privacy concerns, cumbersome data processing procedures, and compatibility issues due to varying formats and protocols across different medical devices.
Innovation Solution
An apparatus and method for extracting medical information using a processor and memory to detect user inputs, classify feature data, specify regions of interest, and perform medical analysis using pre-trained models, enabling seamless integration and efficient extraction of medical data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If medical information is managed through PACS, then medical information can be stored and accessed, but patient privacy is compromised and processing procedures become cumbersome
Solution Approach 1:
The patent extracts only the necessary medical information (specific features and attributes) from the complete medical data set stored in PACS, rather than transferring or processing the entire data set. This extraction approach protects patient privacy by minimizing data exposure while still enabling the required processing and analysis functions.
Solution Approach 2:
The patent segments medical information into distinct components including patient information, examination information, image data, and extracted features. By dividing the data structure into manageable segments with controlled access, the system enables selective processing of specific segments while maintaining privacy protection for sensitive portions.
2Adaptability or versatility
If medical information from various medical devices is integrated, then comprehensive medical data is available, but compatibility between different devices is insufficient
Solution Approach 1:
The patent implements a universal extraction framework that can process medical information from multiple different medical devices and imaging modalities through a common interface. The system extracts standardized features (such as CNN extraction results, classification results, and segmentation results) that are device-agnostic, enabling comprehensive data integration without requiring device-specific processing for each source.
Solution Approach 2:
The patent transforms medical information from various devices into a standardized parameter format, converting different device protocols and formats into unified feature representations. This parameter standardization includes normalizing image attributes, feature data structures, and metadata formats, thereby resolving compatibility issues while maintaining the ability to handle diverse input sources.
3Reliability
If multiple steps are used for uploading and downloading medical information, then data security is maintained, but processing efficiency is reduced
Solution Approach 1:
The patent performs preliminary extraction of medical information features directly at the source system before data transfer or storage operations. By pre-processing and extracting essential features in advance, the system reduces the amount of data that needs to be uploaded and downloaded in subsequent steps, thereby improving processing efficiency while maintaining security through selective data handling.
Solution Approach 2:
The patent introduces an intermediary extraction module that acts as a mediator between the PACS system and the processing system. This intermediary extracts and transforms only the necessary feature data, reducing the number and complexity of upload/download steps while maintaining data security through controlled, selective data transfer rather than bulk operations.
Data Source
AI summary
Disclosed are an apparatus and method for extracting medical information, according to an embodiment. The apparatus for extracting medical information, according to an embodiment, comprises: one or more processors; and a memory that stores instructions for executing the one or more processors, wherein the processors: detect a user input requesting extraction of medical information of a target image, which includes attribute data of the target image and feature data of the target image; extract the attribute data of the target image in response to the user input; extract the medical information by classifying the feature data of the target image; and specify, by using a pre-trained model, a region of interest within the target image, which corresponds to the target image and the medical information.