Apparatus for predicting metadata of medical image and method thereof

Abstract

This disclosure relates to a computerized method to perform a machine learning on a relationship between medical images and metadata using a neural network and acquiring metadata by applying a machine learning model to medical images, and a method thereof. The apparatus and method may include training a prediction model for predicting metadata of medical images based on multiple medical images for learning and metadata matched with each of multiple medical images and predicting metadata of input medical image.

What is claimed is: 
     
         1 . A computerized medical image analysis method, using a hardware processor and a hardware memory, comprising:
 processing, using a prediction model machine-trained for predicting metadata of a medical image, an input medical image to obtain predicted metadata of the input medical image, wherein the input medical image is obtained from an input digital imaging and communications in medicine (DICOM) file comprising an input DICOM header of the input medical image, and wherein the predicted metadata comprises at least one of information related to one or more objects included in the input medical image, information about a shooting environment of the input medical image, or information related to a display method of the input medical image;   adjusting the input medical image using the predicted metadata; and   processing the adjusted input medical image to locate a lesion contained in the adjusted input medical image.   
     
     
         2 . The method of  claim 1 , wherein the information related to one or more objects included in the input medical image comprises at least one of information about one or more body parts included in the input medical image or information about a patient of the input medical image,
 wherein the shooting environment information comprises at least one of modality information of the input medical image or information about a shooting method of the input medical image, and   wherein the display method information comprises at least one of window center information, window width information, color inversion information, image rotation information, or image flip information of the input medical image.   
     
     
         3 . The method of  claim 2 , further comprising adjusting a brightness or contrast of the input medical image based on at least one of the window center information or the window width information. 
     
     
         4 . The method of  claim 2 , wherein the window width information represents a difference between a minimum pixel value and a maximum pixel value of the input medical image. 
     
     
         5 . The method of  claim 2 , wherein the window center information represents at least one of a first threshold value of the input medical image greater than a minimum pixel value, or a second threshold value of the input medical image less than a maximum pixel value. 
     
     
         6 . The method of  claim 1 , further comprising matching and saving the predicted metadata of the input medical image in association with the input medical image. 
     
     
         7 . The method of  claim 6 , wherein the predicted metadata is saved on the input DICOM header of the input medical image. 
     
     
         8 . The method of  claim 1 , wherein adjusting the input medical image comprises:
 adjusting at least one of a window center, a window width, a color, or an output direction of the input medical image based on the predicted metadata.   
     
     
         9 . The method of  claim 1 , wherein the predicted metadata corresponds to at least one of a plurality of elements of the input DICOM header. 
     
     
         10 . The method of  claim 1 , further comprising:
 verifying whether body part information of the predicted metadata matches a target body part for which abnormality is to be detected;   verifying whether modality information of the predicted metadata is appropriate for detecting anomaly; and   verifying whether patient information of the predicted metadata is appropriate for detecting abnormality.   
     
     
         11 . A medical image analysis apparatus comprising a memory storing computer-executable instructions and a processor configured to execute the computer-executable instructions,
 wherein the processor is configured, by executing the computer-executable instructions, to perform:   processing, using a prediction model machine-trained for predicting metadata of a medical image, an input medical image to obtain predicted metadata of the input medical image, wherein the input medical image is obtained from an input digital imaging and communications in medicine (DICOM) file comprising an input DICOM header of the input medical image, and wherein the predicted metadata comprises at least one of information related to one or more objects included in the input medical image, information about a shooting environment of the input medical image, or information related to a display method of the input medical image;   adjusting the input medical image using the predicted metadata; and   processing the adjusted input medical image to locate a lesion contained in the adjusted input medical image.   
     
     
         12 . The apparatus of  claim 11 , wherein the information related to the one or more objects included in the input medical images comprises at least one of information about one or more body parts included in the input medical image, or information about a patient of the input medical image,
 wherein the shooting environment comprises at least one of modality information of the input medical image, or information about a shooting method of the input medical image, and   wherein the display method information comprises at least one of window center information, window width information, color inversion information, image rotation information, or image flip information of the input medical image.   
     
     
         13 . The apparatus of  claim 12 , wherein the processor is further configured to adjust a brightness or contrast of the input medical image based on at least one of the window center information, or the window width information. 
     
     
         14 . The apparatus of  claim 12 , wherein the window width represents a difference between a minimum pixel value and a maximum pixel value of the input medical image. 
     
     
         15 . The apparatus of  claim 12 , wherein the window center represents at least one of a first threshold value of the input medical image greater than a minimum pixel value, or a second threshold value of the input medical image less than a maximum pixel value. 
     
     
         16 . The apparatus of  claim 11 , wherein the processor is further configured, by executing the computer-executable instructions, to perform matching and saving the predicted metadata of the input medical image with the input medical image. 
     
     
         17 . The apparatus of  claim 16 , wherein the processor is further configured, by executing the computer-executable instructions, to perform saving the predicted metadata on the input DICOM header of the input medical image. 
     
     
         18 . The apparatus of  claim 11 , wherein the processor is further configured, by executing the computer-executable instructions, to perform adjusting at least one of a window center, a window width, a color, or an output direction of the input medical image based on the predicted metadata. 
     
     
         19 . The apparatus of  claim 11 , wherein the predicted metadata corresponds to at least one of a plurality of elements of the input DICOM header. 
     
     
         20 . The apparatus of  claim 11 , wherein the processor is further configured to:
 verify whether body part information of the predicted metadata matches a target body part for which abnormality is to be detected;   verify whether modality information of the predicted metadata is appropriate for detecting anomaly; and   verify whether patient information of the predicted metadata is appropriate for detecting abnormality.