Image-based aperture verification system for multi-leaf collimator

Patent metadata

Field	Value
Publication number	US-9950193-B2
Application number	US-201615008232-A
Country	US
Kind code	B2
Filing date	Jan 27, 2016
Priority date	Jun 12, 2015
Publication date	Apr 24, 2018
Grant date	Apr 24, 2018

Abstract

A radiation delivery system includes a multi-leaf collimator (MLC) that includes a housing, a plurality of leaves disposed within the housing, and an image sensor disposed within the housing at a position that is offset from a beam axis. The plurality of leaves are movable to define an aperture for the MLC, and the image sensor is directed toward the plurality of leaves. The image sensor is to generate an oblique-view image of the aperture. The radiation delivery system additionally includes a processing device to receive the oblique-view image, transform the oblique-view image into a top-view image having a reference coordinate space, and determine whether the aperture for the MLC corresponds to a specified aperture based on the top-view image.

First claim

What is claimed is:

1. A radiation delivery system comprising:
a multi-leaf collimator (MLC) comprising:
a housing;
a plurality of leaves disposed within the housing, wherein the plurality of leaves are movable to define an aperture for the MLC; and
an image sensor disposed within the housing at a position that is offset from a beam axis, wherein the image sensor is directed toward the plurality of leaves, the image sensor to generate an oblique-view image of the aperture; and

a processing device to:
receive the oblique-view image;
transform the oblique-view image into a top-view image having a reference coordinate space, wherein to transform the oblique-view image the processing device is to apply a geometric transformation to the oblique-view image, wherein the oblique-view image to be generated by the image sensor is a distorted oblique-view image comprising optical lens distortion, and wherein the processing device is further to perform an additional transformation to remove the optical lens distortion from the distorted oblique-view image prior to transformation of the oblique-view image into the top-view image; and
determine whether the aperture for the MLC corresponds to a specified aperture based on the top-view image;

wherein the processing device is to generate a transformation matrix that represents the geometric transformation, wherein to generate the transformation matrix the processing device is to:
cause the plurality of leaves to define a predefined calibration aperture that corresponds to a known aperture shown in a template image;
cause the image sensor to generate an additional oblique-view image of the calibration aperture;
perform image processing on the additional oblique-view image of the calibration aperture to find first keypoints;
perform image processing on the template image to find second keypoints;
perform feature matching to match members of the first keypoints to corresponding members of the second keypoints; and
determine transformations to align the first keypoints with the second keypoints.

2. The radiation delivery system of claim 1 , wherein the plurality of leaves follow parallel linear trajectories, and wherein changes in position of each of the plurality of leaves are shown as changes in position on a single axis of the reference coordinate space in the top-view image.

3. The radiation delivery system of claim 1 , wherein for the known aperture the plurality of leaves are in non-uniform positions and cause a plurality of keypoints.

4. The radiation delivery system of claim 1 , wherein:
the plurality of leaves comprises a first bank of leaves disposed at a first side of the beam axis and a second bank of leaves disposed at a second side of the beam axis that is opposite to the first side of the beam axis.

5. The radiation delivery system of claim 1 , wherein to determine whether the aperture for the MLC corresponds to a specified aperture based on the top-view image, the processing device is to:
determine leaf positions of the plurality of leaves from the top-view image;
compare the leaf positions from the top-view image to additional leaf positions associated with the specified aperture to determine a variation between the leaf positions and the additional leaf positions;
determine that the variation is less than a threshold value; and
determine that there is a match between the leaf positions and the additional leaf positions.

6. The radiation delivery system of claim 1 , wherein the oblique-view image comprises a keystone effect, and wherein the top-view image emulates a view of the plurality of leaves that would be generated if the image sensor was oriented parallel to the beam axis and positioned approximately on the beam axis.

7. The radiation delivery system of claim 1 , further comprising:
an imaging device that comprises the image sensor and a sensor control board, wherein the sensor control board is separated from the image sensor and is positioned away from the beam axis.

8. A method comprising:
generating an oblique-view image of an aperture for a multi-leaf collimator (MLC) by an image sensor disposed within the MLC, wherein the image sensor is offset from a beam axis of the MLC, and wherein the aperture comprises a leaf position for each of a plurality of leaves of the MLC;
transforming the oblique-view image into a top-view image having a reference coordinate space by applying a geometric transformation to the distorted oblique-view image;
generating a transformation matrix that represents the geometric transformation by:
causing the MLC to have a predefined calibration aperture that corresponds to a known aperture shown in a template image;
generating an additional oblique-view image of the calibration aperture;
performing image processing on the additional oblique-view image to find first keypoints;
performing image processing on the template image to find second keypoints;
performing feature matching to match members of the first keypoints to corresponding members of the second keypoints; and
determining transformations to align the first keypoints with the second keypoints; and

determine whether the aperture for the MLC corresponds to a specified aperture based on the top-view image.

9. The method of claim 8 , wherein the plurality of leaves follows parallel linear trajectories, and wherein changes in position of each of the plurality of leaves are shown as changes in position on a single axis of the reference coordinate space in the top-view image.

10. The method of claim 8 , wherein the oblique-view image is a distorted oblique-view image comprising optical lens distortion, the method further comprising:
performing an additional transformation to remove the optical lens distortion from the distorted oblique-view image prior to transforming the oblique-view image into the top-view image.

11. The method of claim 8 , wherein for the known aperture the plurality of leaves are in non-uniform positions and cause a plurality of keypoints.

12. The method of claim 8 , further comprising:
periodically generating the transformation matrix to confirm calibration of the MLC.

13. The method of claim 8 , wherein the oblique-view image comprises a keystone effect, and wherein the top-view image emulates a view of the plurality of leaves that would be generated if the image sensor was oriented orthogonal to the beam axis and its center was positioned approximately on the beam axis.

14. The method of claim 8 , wherein determining whether the aperture for the MLC corresponds to a specified aperture based on the top-view image comprises:
determining leaf positions of the plurality of leaves from the top-view image;
comparing the leaf positions from the top-view image to additional leaf positions associated with the specified aperture to determine a variation between the leaf positions and the additional leaf positions;
determining that the variation is less than a threshold value; and
determining that there is a match between the leaf positions and the additional leaf positions.

15. The method of claim 8 , wherein determining whether the aperture for the MLC corresponds to the specified aperture comprises:
comparing the aperture to a planned aperture specified in a treatment plan.

16. The method of claim 8 , wherein determining whether the aperture for the MLC corresponds to the specified aperture comprises:
comparing the aperture to an additional aperture determined by an additional aperture verification system.

Assignees

• Accuray Inc

Inventors

• Chappelow Jonathan Cecil
• Jordan Petr
• Maurer Calvin
• Singh Sarbjit
• Fitting Andrew W
• Myronenko Andriy
• Gao Jian

Classifications

• H04N23/51

  Housings · CPC title

• H04N23/57

  Mechanical or electrical details of cameras or camera modules specially adapted for being embedded in other devices · CPC title

• G21K1/046

  varying the contour of the field, e.g. multileaf collimators · CPC title

• A61N5/1067

  in real time, i.e. during treatment · CPC title

• H04N5/2252

  Electricity · mapped topic