Apparatus, method, and program for training discriminator discriminating disease region, discriminator discriminating disease region, disease region discrimination apparatus, and disease region discrimination program
地域：
Tokyo
摘要
A discriminator includes a common learning unit and a plurality of learning units that are connected to an output unit of the common learning unit. The discriminator is trained, using a plurality of data sets of a first image obtained by capturing an image of a subject that has developed a disease and an image data of a disease region in the first image, such that information indicating the disease region is output from a first learning unit in a case in which the first image is input to the common learning unit. In addition, the discriminator is trained, using a plurality of data sets of an image set obtained by registration between the first image and a second image whose type is different from the type of the first image, such that an estimated image of the second image is output from an output unit of a second learning unit.
說明書
The present application claims priority under 35 U.S.C. § 119 to Japanese Patent Application No. 2018-186790, filed on Oct. 1, 2018. Each of the above application is hereby expressly incorporated by reference, in its entirety, into the present application.
The present disclosure relates to an apparatus, method, and a non-transitory computer readable recording medium storing a program for training a discriminator discriminating a disease region, a discriminator discriminating a disease region, a disease region discrimination apparatus, and a non-transitory computer readable recording medium storing a disease region discrimination program.
In recent years, advances in medical apparatuses, such as a computed tomography (CT) apparatus and a magnetic resonance imaging (MRI) apparatus, have made it possible to perform image diagnosis using high-resolution medical images with higher quality. In particular, in a case in which a target part is the brain, image diagnosis using, for example, CT images and MR images makes it possible to specify regions causing cerebrovascular disorders, such as cerebral infarction and cerebral hemorrhage. Therefore, various methods for supporting image diagnosis have been proposed.
權利要求
What is claimed is:1. A method training a discriminator which is implemented by an electronic device having a hardware processor and comprising a common learning unit that includes an input unit and an output unit and a plurality of learning units each of which includes an input unit which is connected to the output unit of the common learning unit and an output unit, the method comprising:training the discriminator by the hardware processor, using a plurality of data sets of a first image obtained by capturing an image of a subject that has developed a disease and an image data of a disease region in which the disease appears in the first image, such that information indicating the disease region is output from the output unit of a first learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit; and training the discriminator by the hardware processor, using a plurality of data sets of an image set obtained by registration between the first image and a second image which is obtained by capturing an image of the same subject as described above and whose type is different from a type of the first image, such that an estimated image of the second image is output from the output unit of a second learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit. 2. The learning method according to claim 1 ,wherein the common learning unit is trained by the hardware processor, using the plurality of data sets of a first image and an image data of the disease region in which the disease appears in the first image and the plurality of data sets of an image set obtained by registration between the first image and a second image, such that a feature amount data of the medical image is output from the output unit of the common learning unit in a case in which the medical image is input to the input unit of the common learning unit. 3. The learning method according to claim 1 ,wherein the first image and the second image are captured under different imaging conditions. 4. The learning method according to claim 1 ,wherein the first image and the second image are captured by different imaging principles. 5. The learning method according to claim 1 ,wherein the first image is a computed tomography (CT) image and the second image is a magnetic resonance (MR) image. 6. The learning method according to claim 5 ,wherein the MR image is a diffusion-weighted image. 7. The learning method according to claim 1 ,wherein the subject is a brain of a patient that has developed cerebral infarction, and the disease region is an infarction region. 8. The learning method according to claim 2 ,wherein the subject is a brain of a patient that has developed cerebral infarction, and the disease region is an infarction region. 9. The learning method according to claim 5 ,wherein the subject is a brain of a patient that has developed cerebral infarction, and the disease region is an infarction region. 10. The learning method according to claim 1 ,wherein each of the common learning unit and the plurality of learning units is a neural network that comprises an input layer as the input unit, a plurality of intermediate layers, and an output layer as the output unit. 11. A apparatus training a discriminator which is implemented by an electronic device having a hardware processor and comprising a common learning unit that includes an input unit and an output unit and a plurality of learning units each of which includes an input unit which is connected to the output unit of the common learning unit and an output unit, the learning apparatus configured to:train the discriminator by the hardware processor, using a plurality of data sets of a first image obtained by capturing an image of a subject that has developed a disease and an image data of a disease region in which the disease appears in the first image, such that information indicating the disease region is output from the output unit of a first learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit; and train the discriminator by the hardware processor, using a plurality of data sets of an image set obtained by registration between the first image and a second image which is obtained by capturing an image of the same subject as described above and whose type is different from a type of the first image, such that an estimated image of the second image is output from the output unit of a second learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit. 12. A learning apparatus according to claim 11 , further configured to:train the common learning unit by the hardware processor, using the plurality of data sets of a first image obtained by capturing an image of a subject that has developed the disease and an image data of the disease region in which the disease appears in the first image and the plurality of data sets of an image set obtained by registration between the first image and a second image, such that a feature amount data of the medical image is output from the output unit of the common learning unit in a case in which the medical image is input to the input unit of the common learning unit. 13. A non-transitory computer readable medium for storing a learning program training a discriminator which is implemented by an electronic device having a hardware processor and comprising a common learning unit that includes an input unit and an output unit and a plurality of learning units each of which includes an input unit which is connected to the output unit of the common learning unit and an output unit, the learning program causing a computer to perform:a process of training the discriminator by a hardware processor, using a plurality of data sets of a first image obtained by capturing an image of a subject that has developed a disease and an image data of a disease region in which the disease appears in the first image, such that information indicating the disease region is output from the output unit of a first learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit; and a process of training the discriminator by the hardware processor, using a plurality of data sets of an image set obtained by registration between the first image and a second image which is obtained by capturing an image of the same subject as described above and whose type is different from a type of the first image, such that an estimated image of the second image is output from the output unit of a second learning unit among the plurality of learning units in a case in which the first image is input to the input unit of the common learning unit. 14. A non-transitory computer readable medium for storing a learning program according to claim 13 , the learning program further causing a computer to perform:a process of training the common learning unit by the hardware processor, using the plurality of data sets of a first image obtained by capturing an image of a subject that has developed the disease and an image data of the disease region in which the disease appears in the first image and the plurality of data sets of an image set obtained by registration between the first image and a second image, such that a feature amount data of the medical image is output from the output unit of the common learning unit in a case in which the medical image is input to the input unit of the common learning unit. 15. A discriminator that is trained by the learning method according to claim 1 . 16. A discriminator that is trained by the learning apparatus according to claim 11 . 17. A discriminator that is trained by the learning program according to claim 13 . 18. A disease region discrimination apparatus comprising:a processor configured to acquire a first image which is a discrimination target; and the discriminator according to claim 15 that discriminates a disease region in the first image which is the discrimination target. 19. The disease region discrimination apparatus according to claim 18 ,the processor displays a discrimination result of the discriminator on a displayer. 20. A non-transitory computer readable medium for storing a disease region discrimination program that causes a computer to perform:a process of acquiring a first image which is a discrimination target; and a process of allowing the discriminator according to claim 15 to discriminate a disease region in the first image which is the discrimination target.
意見反饋