Medical image processing apparatus and medical image processing method
AssigneeFUJIFILM Healthcare Corporation
InventorsHisashi Takahashi, Taiga Gotou
Life of this patent
Estimating the term…
Abstract
A medical image processing apparatus reducing noise of a medical image acquired includes a smoothing unit forming a smoothed image of the medical image, a route forming unit forming a route that is a pixel group positioned continuously in the smoothed image and fulfills a route condition, and a noise reducing unit extracting a pixel group corresponding to the route in the medical image and reducing noise of the medical image based on the extracted pixel group.
Claims9
A medical image processing apparatus reducing noise of a medical image acquired, the medical image processing apparatus comprising:
- a smoothing unit forming a smoothed image of the medical image, wherein
- the smoothed image is formed by at least one of:
- smoothing the medical image by calculating one of an average pixel value and a median pixel value of each of an object pixel and an adjacent pixel group, and substituting the pixel value of the object pixel with the calculated average value or the calculated median value,
- using a noise reducing method by a successive approximation method or by utilizing artificial intelligence, and
- acquiring smoothed projection data by smoothing projection data that is used in formation of the medical image and subjecting the smoothed projection data to reconstruction processing;
- a route forming unit forming a route that is a first pixel group having coordinates positioned continuously in the smoothed image and fulfills a route condition, wherein the route forming unit forms the first pixel group from an originating pixel to a newest watched pixel as the route by repeating setting a pixel where a difference from a pixel value of a watched pixel is minimum from a third pixel group adjacent to the watched pixel as a new watched pixel until the route condition is no longer fulfilled after the originating pixel selected from the smoothed image is set as the watched pixel;
- and
- a noise reducing unit configured to:
- extract a second pixel group from the medical image corresponding to the coordinates of the first pixel group by calculating a representative pixel value of the coordinates;
- and
- reduce noise of the medical image based on the extracted second pixel group by substituting a pixel value in the medical image with the representative pixel value.
A medical image processing method reducing noise of a medical image acquired, the medical image processing method comprising:
- a smoothing step of forming a smoothed image of the medical image wherein,
- the smoothed image is formed by at least one of:
- smoothing the medical image by calculating one of an average pixel value and a median pixel value of each of an object pixel and an adjacent pixel group, and substituting the pixel value of the object pixel with the calculated average value or the calculated median value,
- using a noise reducing method by a successive approximation method or by utilizing artificial intelligence, and
- acquiring smoothed projection data by smoothing projection data that is used in formation of the medical image and subjecting the smoothed projection data to reconstruction processing;
- a route forming step of forming a route that is a first pixel group positioned continuously in the smoothed image and fulfills a route condition, wherein the route forming unit forms the first pixel group from an originating pixel to a newest watched pixel as the route by repeating setting a pixel where a difference from a pixel value of a watched pixel is minimum from a third pixel group adjacent to the watched pixel as a new watched pixel until the route condition is no longer fulfilled after the originating pixel selected from the smoothed image is set as the watched pixel;
- and
- a noise reducing step of:
- extracting a second pixel group from the medical image corresponding to the coordinates of the first pixel group by calculating a representative pixel value of the coordinates and;
- reducing noise of the medical image based on the extracted second pixel group by substituting a pixel value in the medical image with the representative pixel value.
Dependencies are inferred from the claim text.