Apparatus and method for processing medical image using predicted metadata
Inventors
Park, Jong Chan • Yoo, Dong Geun • YOU, Ki Hyun • Nam, Hyeon Seob • Lee, Hyun Jae • Lee, Sang Hyup
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12136483-B2
Publication Date
2024-11-05
Expiration Date
Abstract
The present disclosure relates to a medical image analysis method using a processor and a memory which are hardware. The method includes generating predicted second metadata for a medical image by using a prediction model, and determining a processing method of the medical image based on one of first metadata stored corresponding to the medical image and the second metadata.
Core Innovation
The disclosure relates to a medical image analysis apparatus that receives a medical image and obtains metadata based on the medical image. The apparatus determines, based on the metadata, whether the medical image is suitable for analysis by a machine learning model configured to detect abnormality, and determines that the medical image is not suitable in response to information related to at least one item included in the metadata not satisfying a predetermined condition.
In response to the information related to the at least one item included in the metadata not satisfying the predetermined condition, the operations further comprise obtaining a new medical image of a patient corresponding to the medical image, or performing an operation for obtaining the new medical image of the patient. The disclosure further specifies obtaining metadata by generating first predicted metadata with a prediction model and selecting either the first predicted metadata or second stored metadata corresponding to the medical image as the metadata.
The disclosure also describes determining a reference value related to determination of a machine learning model configured to detect abnormality based on the metadata, obtaining result information by applying the medical image to the machine learning model, and obtaining final result information by comparing the reference value with the result information. The comparison-based determination yields final result information based on the relationship between the reference value and the result information, thereby supporting abnormality detection.
Claims Coverage
The document includes multiple independent claims covering metadata-based suitability determination for abnormality detection, reference-value comparison between model result information and a metadata-dependent reference value, and corresponding implementations for an apparatus, a method, and a non-transitory computer-readable medium. Across the independent claims, inventive features include conditional suitability using predetermined metadata conditions, metadata acquisition via prediction plus selection of predicted vs stored metadata, and comparison-based final determination using a reference value tied to metadata.
Metadata-based suitability for abnormality model analysis
Determining, based on the metadata, whether the medical image is suitable for analysis by a machine learning model configured to detect abnormality, wherein the determining comprises determining that the medical image is not suitable for the analysis by the machine learning model in response to information related to at least one item included in the metadata not satisfying a predetermined condition.
Obtaining a new medical image in response to metadata unsuitability
Obtaining a new medical image of a patient corresponding to the medical image, or performing an operation for obtaining the new medical image of the patient, in response to the information related to at least one item included in the metadata not satisfying the predetermined condition.
Prediction-model metadata acquisition with selection between predicted and stored metadata
Generating first predicted metadata with a prediction model and selecting either the first predicted metadata or second stored metadata corresponding to the medical image as the metadata.
Reference value determination for abnormality detection based on metadata
Determining a reference value related to determination of a machine learning model configured to detect abnormality based on the metadata.
Comparison-based final result information from reference value and result information
Obtaining result information by applying the medical image to the machine learning model, and obtaining final result information by comparing the reference value with the result information.
Overall, the independent claims are directed to using metadata items and predetermined conditions to determine suitability for abnormality detection by a machine learning model, including obtaining a new medical image when unsuitability is detected, determining a metadata-related reference value and producing final result information by comparing the reference value with model result information, and acquiring metadata using a prediction model with selection between predicted metadata and stored metadata.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?