Device and method for providing sinus image and learning method thereof
An apparatus for providing a maxillary sinus image according to one embodiment includes a first artificial neural network model configured to distinguish a partial coronal plane region-of-interest from a coronal plane image of a head, a second artificial neural network model configured to, relative to an axial plane image of the head, extract a two-dimensional maxillary sinus image in an axial plane region-of-interest that corresponds to a position of the coronal plane region-of-interest, and an image processor configured to generate a three-dimensional maxillary sinus image using the two-dimensional maxillary sinus image.
1 . An apparatus for providing a maxillary sinus image, the apparatus comprising:
a first artificial neural network model configured to distinguish a partial coronal plane region-of-interest from a coronal plane image of a head and to output coordinate information of the coronal plane region-of-interest;
an image processor configured to, based on the coordinate information of the coronal plane region-of-interest, use the coordinate information as coordinate information of an axial plane region-of-interest corresponding to the coronal plane region-of-interest, and distinguish the axial plane region-of-interest from an axial plane image of the head;
a second artificial neural network model configured to extract a two-dimensional maxillary sinus image in the axial plane region-of-interest that corresponds to a position of the coronal plane region-of-interest; and
the image processor further configured to generate a three-dimensional maxillary sinus image using the two-dimensional maxillary sinus image.
2 . The apparatus of claim 1 , wherein:
the image processor provides a cropped image, which includes the axial plane region-of-interest in the axial plane image of the head, to the second artificial neural network model, based on three-dimensional coordinate values of the coronal plane region-of-interest; and
the second artificial neural network model extracts the two-dimensional maxillary sinus image from the cropped image.
3 . The apparatus of claim 2 , wherein:
the coronal plane image of the head and the axial plane image of the head are images captured by radiation imaging apparatuses that use the same sagittal-axial-coronal coordinate system;
the first artificial neural network model outputs sagittal-axial-coronal coordinate information of the coronal plane region-of-interest distinguished from the coronal plane image of the head; and
the image processor obtains sagittal plane coordinate values and axial plane coordinate values of the axial plane region-of-interest from sagittal plane coordinate values and coronal plane coordinate values of the coronal plane region-of-interest and distinguishes the axial plane region-of-interest from the axial plane image of the head.
4 . The apparatus of claim 2 , wherein the second artificial neural network model learns using a cropped image including the axial plane region-of-interest as learning input data and using a masking image, in which a maxillary sinus region in the cropped image is masked, as label data.
5 . The apparatus of claim 1 , wherein the first artificial neural network model learns using a learning coronal plane image of the head as learning input data and using a masking image, in which the coronal plane region-of-interest in the learning coronal plane image of the head is combined, as label data.
6 . The apparatus of claim 1 , wherein the first artificial neural network model extracts a plurality of candidate regions that are distinguishable as the coronal plane region-of-interest from the coronal plane image of the head and, according to a result of comparison of positions based on coronal plane coordinate values of the plurality of candidate regions or comparison of positions based on a central axis of the axial plane, determines at least one candidate region as the coronal plane region-of-interest.
7 . A method for providing a maxillary sinus image that is performed by an apparatus for providing a maxillary sinus image, the method comprising:
distinguishing a partial coronal plane region-of-interest from an input coronal plane image of a head using a first artificial neural network model and outputting coordinate information of the coronal plane region-of-interest;
based on the coordinate information of the coronal plane region-of-interest, using the coordinate information as coordinate information of an axial plane region-of-interest corresponding to the coronal plane region-of-interest, and distinguishing the axial plane region-of-interest from an axial plane image of the head;
extracting, by a second artificial neural network model, a two-dimensional maxillary sinus image in the axial plane region-of-interest that corresponds to a position of the coronal plane region-of-interest; and
generating a three-dimensional maxillary sinus image using the two-dimensional maxillary sinus image.
8 . A non-transitory computer-readable recording medium storing a computer program, wherein the computer program includes instructions that, upon being executed by a processor, allow the processor to perform a method for providing a maxillary sinus image, the method comprising:
distinguishing a partial coronal plane region-of-interest from an input coronal plane image of a head using a first artificial neural network model and outputting coordinate information of the coronal plane region-of-interest;
based on the coordinate information of the coronal plane region-of-interest, using the coordinate information as coordinate information of an axial plane region-of-interest corresponding to the coronal plane region-of-interest, and distinguishing the axial plane region-of-interest from an axial plane image of the head;
extracting, by a second artificial neural network model, a two-dimensional maxillary sinus image in the axial plane region-of-interest that corresponds to a position of the coronal plane region-of-interest; and
generating a three-dimensional maxillary sinus image using the two-dimensional maxillary sinus image.
9 . A learning method of an apparatus for providing a maxillary sinus image, the learning method comprising:
to distinguish a partial coronal plane region-of-interest from an input coronal plane image of a head, causing a first artificial neural network model to learn using a learning coronal plane image of the head as first learning input data and using a masking image, in which the coronal plane region-of-interest in the learning coronal plane image of the head is masked, as first label data, and to output coordinate information of the coronal plane region-of-interest; and
to, based on the coordinate information of the coronal plane region-of-interest, use the coordinate information as coordinate information of an axial plane region-of-interest corresponding to the coronal plane region-of-interest and distinguish the axial plane region-of-interest from an axial plane image of the head, and to extract a two-dimensional maxillary sinus image in a cropped image including the axial plane region-of-interest that corresponds to a position of the coronal plane region-of-interest in an axial plane image of the head, causing a second artificial neural network model to learn using a learning cropped image as second learning input data and using a masking image, in which a maxillary sinus region in the cropped image is masked, as second label data.