Medical image processing apparatus and medical image processing system
A medical image processing apparatus according to an embodiment comprises a memory and processing circuitry. The memory is configured to store a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site. The processing circuitry is configured to process first data into second data using, among the neural networks, the neural network corresponding to the imaging target site for the first data, wherein the first data is input to the input layer and the second data is output from the output layer.
1. A medical image processing apparatus comprising:
a memory storing a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site; and
processing circuitry configured to process a plurality of sets of first data acquired using respective different field-of-view (FOV) sizes corresponding to one of the plurality of imaging target sites into respective sets of second data each of which has a granularity level independent of the FOV size, each of the sets of first data being processed using a corresponding one of the plurality of neural networks,
wherein for the one of the plurality of neural networks corresponding to the imaging target site, the corresponding set of the first data is input to the input layer and the corresponding set of the second data is output from the output layer.
2. The medical image processing apparatus according to claim 1 , wherein
the memory is configured to store a plurality of neural networks corresponding to a plurality of imaging conditions, respectively, the neural networks corresponding to the respective imaging conditions each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired under imaging conditions equal to or similar to the corresponding imaging condition; and
the processing circuitry is configured to process the second data into third data using, among the neural networks corresponding to the respective imaging conditions, the neural network corresponding to the imaging condition for the second data, wherein the second data is input to the input layer and the third data is output from the output layer.
3. The medical image processing apparatus according to claim 1 , wherein
the memory is configured to store a plurality of neural networks corresponding to a plurality of noise levels, respectively, the neural networks corresponding to the respective noise levels each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets having noise levels equal to or similar to the corresponding noise level; and
the processing circuitry is configured to process the second data into third data using, among the neural networks corresponding to the respective noise levels, the neural network corresponding to the noise level of the second data, wherein the second data is input to the input layer and the third data is output from the output layer.
4. The medical image processing apparatus according claim 1 , wherein the first data is data before reconstruction.
5. The medical image processing apparatus according to claim 1 , wherein the first data is reconstructed image data.
6. The medical image processing apparatus according to claim 5 , wherein
the memory is configured to store a plurality of neural networks corresponding to the FOV sizes, respectively, the neural networks corresponding to the respective FOV sizes each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding FOV size, and
the processing circuitry is configured to process the second data into third data using, among the neural networks corresponding to the respective FOV sizes, the neural network corresponding to the FOV size of the second data, wherein the second data is input to the input layer and the third data is output from the output layer.
7. The medical image processing apparatus according to claim 5 , wherein
the first data comprises data corresponding to a first slice plane, data corresponding to a second slice place, and data corresponding to a third slice plane, the second slice plane and the third slice plane adjacent the first slice plane with respect to a slice direction, and
the neural networks are convolutional neural networks each using a kernel with channels corresponding to the first slice plane, the second slice plane, and the third slice plane.
8. The medical image processing apparatus according to claim 1 , wherein the multiple data sets for the learning processing comprise multiple pre-denoised data sets, and multiple denoised data sets corresponding to the respective pre-denoised data sets.
9. The medical image processing apparatus according to claim 1 , wherein the multiple data sets for the learning processing comprise multiple pre-artifact removed data sets, and multiple artifact removed data sets corresponding to the respective pre-artifact removed data sets.
10. The medical image processing apparatus according to claim 1 , wherein the multiple data sets for the learning processing comprise multiple pre-noise granularity processed data sets, and multiple noise granularity processed data sets corresponding to the respective pre-noise granularity processed data sets.
11. The medical image processing apparatus according to claim 1 , wherein the processing circuitry is configured to perfoiiii noise reduction processing for the first data or the second data based on noise estimated using a geometry of an imaging system and noise estimated using a statistical noise model.
12. The medical image processing apparatus according to claim 1 , wherein the processing circuitry is configured to use three dimensional information to retain a sharp structure in the first data or the second data and selectively remove noise.
13. A medical image processing system comprising a server apparatus and a client apparatus, wherein
the server apparatus comprises:
a memory storing a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site; and
processing circuitry configured to process a plurality of sets of first data acquired using respective different field-of-view (FOV) sizes corresponding to one of the plurality of imaging target sites into respective sets of second data each of which having a granularity level independent of the FOV size using one of the plurality of neural networks, wherein for the one of the plurality of neural networks corresponding to the imaging target site, the corresponding set of the first data is input to the input layer and the corresponding set of the second data is output from the output layer, and
the client apparatus is configured to receive the second data via a network.
14. The medical image processing apparatus according to claim 1 , wherein the imaging condition includes a size of a field of view, a reconstruction function, or an x-ray dose used for acquiring the first data.
15. The medical image processing apparatus according to claim 1 , wherein
the memory storing the plurality of neural networks including multiple neural networks corresponding to a plurality of FOV sizes for one of the plurality of imaging target sites, and
the processing circuitry is configured to process a set of first data acquired using a certain FOV size into a set of second data, based on one of the multiple neural networks corresponding to the first FOV size.
16. A medical image processing apparatus comprising:
a memory storing a plurality of neural networks corresponding to a plurality of field-of-view (FOV) sizes, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding FOV size; and
processing circuitry configured to process a plurality of sets of first data acquired using respective different FOV sizes into respective sets of second data each of which having a granularity level independent of the FOV size using the neural networks,
wherein for a neural network corresponding to the FOV size, the corresponding set of the first data is input to the input layer and the corresponding set of the second data is output from the output layer.