Medical image processing method
According to an exemplary embodiment of the present disclosure, a medical image processing method performed by a computing device is disclosed. The medical image processing method includes: detecting a region of interest in a medical image by using a pre-trained deep learning model; determining contour information for the region of interest; and generating, based on the contour information, format information defining elements that determine representation of the medical image.
1 . A medical image processing method performed by a computing device including at least one processor, the method comprising:
detecting a region of interest in a medical image by using a pre-trained deep learning model;
determining contour information for the region of interest; and
generating, based on the contour information, format information defining elements that determine representation of the medical image,
wherein the determining of the contour information for the region of interest includes:
identifying a color space of the medical image; and
determining a color of a contour marking the region of interest, based on correlation between colors in a color distribution of the region of interest in the medical image of which the color space is identified.
2 . The method of claim 1 , wherein when the identified color space is an RGB space, the determining of the color of the contour marking the region of interest includes:
deriving a histogram representing a pixel-by-pixel color distribution from the medical image;
determining a first candidate color value based on a frequency of occurrence of colors present in the histogram;
transforming the color space of the medical image into a Hue Saturation Value (HSV) space, and determining a second candidate color value based on the histogram in the HSV space;
determining a third candidate color value in a grayscale based on brightness of pixels included in the medical image; and
determining the color of the contour marking the region of interest based on the first candidate color value, the second candidate color value, and the third candidate color value.
3 . The method of claim 2 , wherein the determining of the first candidate color value based on the frequency of occurrence of the colors present in the histogram includes determining, based on at least one color having a lowest frequency of occurrence among the colors present in the histogram, at least one color value that prominently represents a predetermined color as the first candidate color value.
4 . The method of claim 2 , wherein the determining of the second candidate color value based on the histogram in the HSV space includes:
selecting an unoccupied hue from the histogram in the HSV space;
selecting saturation and a value of brightness at which visual contrast is prominent based on brightness of pixels included in a candidate region including the region of interest; and
determining the second candidate color value based on the selected hue, saturation, and value of brightness.
5 . The method of claim 2 , wherein the determining of the third candidate color value based on the brightness of the pixels included in the medical image includes:
based on the brightness of the pixels included in a candidate region including the region of interest, selecting a color value in a grayscale at which visual contrast is prominent; and
determining the selected color value in the grayscale as the third candidate color value.
6 . The method of claim 1 , wherein when the identified color space is a grayscale space, the determining of the color of the contour marking the region of interest includes:
deriving a histogram representing a pixel-by-pixel color distribution from the medical image;
identifying at least one color not appearing in the region of interest based on the histogram; and
determining, based on the at least one color not appearing in the region of interest, the color of the contour marking the region of interest.
7 . The method of claim 6 , wherein the determining of, based on the at least one color not appearing in the region of interest, the color of the contour marking the region of interest includes determining, based on the at least one color not appearing in the region of interest, colors that are distinguishable from each other in accordance with a type and the number of contours that mark the region of interest.
8 . The method of claim 1 , wherein the determining of the contour information for the region of interest includes determining a representation element related to a shape of the contour of the region of interest based on detection information of the region of interest.
9 . The method of claim 8 , wherein the detection information includes at least one of a probability value regarding presence of the region of interest or a numerical value of the region of interest, and
wherein the representation element related to the shape of the contour of the region of interest includes at least one of a thickness of the contour, sharpness of the contour, precision of the contour, or sharpness of a shadow of the contour.
10 . The method of claim 8 , wherein the determining of the representation element related to the shape of the contour of the region of interest based on the detection information of the region of interest includes determining at least one of a thickness of the contour, sharpness of the contour, precision of the contour, or sharpness of a shadow of the contour, based on a magnitude of at least one of a probability value regarding presence of the region of interest or a numerical value of the region of interest.
11 . The method of claim 10 , wherein as the magnitude of at least one of the probability value regarding the presence of the region of interest or the numerical value of the region of interest increases, a magnitude of at least one of the thickness of the contour, the sharpness of the contour, the precision of the contour, or the sharpness of the shadow of the contour increases.
12 . A medical image processing method performed by a computing device including at least one processor, the method comprising:
receiving format information defining elements that determine representation of a medical image;
combining the medical image with format information corresponding to the medical image; and
generating a user interface reflecting a result of the combination,
wherein the format information is generated based on contour information for a region of interest of the medical image detected using a pre-trained deep learning model,
wherein the determining of the contour information for the region of interest includes:
identifying a color space of the medical image; and
determining a color of a contour marking the region of interest, based on correlation between colors in a color distribution of the region of interest in the medical image of which the color space is identified.
13 . A computing device for processing a medical image, the computing device comprising:
a processor including at least one core;
a memory including program codes executed in the processor; and
a network unit for receiving a medical image,
wherein the processor:
detects a region of interest in the medical image by using a pre-trained deep learning model,
determines contour information for the region of interest, and
generates, based on the contour information, format information defining elements that determine a representation layer of the medical image,
wherein the determining of the contour information for the region of interest includes:
identifying a color space of the medical image; and
determining a color of a contour marking the region of interest, based on correlation between colors in a color distribution of the region of interest in the medical image of which the color space is identified.
14 . The computing device of claim 13 , wherein when the identified color space is an RGB space, the determining of the color of the contour marking the region of interest includes:
deriving a histogram representing a pixel-by-pixel color distribution from the medical image;
determining a first candidate color value based on a frequency of occurrence of colors present in the histogram;
transforming the color space of the medical image into a Hue Saturation Value (HSV) space, and determining a second candidate color value based on the histogram in the HSV space;
determining a third candidate color value in a grayscale based on brightness of pixels included in the medical image; and
determining the color of the contour marking the region of interest based on the first candidate color value, the second candidate color value, and the third candidate color value.
15 . The computing device of claim 14 , wherein the determining of the first candidate color value based on the frequency of occurrence of the colors present in the histogram includes determining, based on at least one color having a lowest frequency of occurrence among the colors present in the histogram, at least one color value that prominently represents a predetermined color as the first candidate color value.
16 . The computing device of claim 14 , wherein the determining of the second candidate color value based on the histogram in the HSV space includes:
selecting an unoccupied hue from the histogram in the HSV space;
selecting saturation and a value of brightness at which visual contrast is prominent based on brightness of pixels included in a candidate region including the region of interest; and
determining the second candidate color value based on the selected hue, saturation, and value of brightness.
17 . The computing device of claim 14 , wherein the determining of the third candidate color value based on the brightness of the pixels included in the medical image includes:
based on the brightness of the pixels included in a candidate region including the region of interest, selecting a color value in a grayscale at which visual contrast is prominent; and
determining the selected color value in the grayscale as the third candidate color value.
18 . The computing device of claim 13 , wherein when the identified color space is a grayscale space, the determining of the color of the contour marking the region of interest includes:
deriving a histogram representing a pixel-by-pixel color distribution from the medical image;
identifying at least one color not appearing in the region of interest based on the histogram; and
determining, based on the at least one color not appearing in the region of interest, the color of the contour marking the region of interest.
19 . The computing device of claim 18 , wherein the determining of, based on the at least one color not appearing in the region of interest, the color of the contour marking the region of interest includes determining, based on the at least one color not appearing in the region of interest, colors that are distinguishable from each other in accordance with a type and the number of contours that mark the region of interest.
20 . The computing device of claim 13 , wherein the determining of the contour information for the region of interest includes determining a representation element related to a shape of the contour of the region of interest based on detection information of the region of interest.