Gaze-tracking driven region of interest segmentation
A region of interest segmentation system includes a display device, a gaze-tracking device, a gaze point collector, a boundary identifier, and a region identifier. The display device displays an image. The gaze-tracking device, generates gaze points relative to the displayed image. The gaze point collector selects gaze points from the generated gaze points corresponding to a region of interest of the displayed image. the boundary identifier estimates a boundary based on the selected gaze points. The region identifier segments the region of interest based on the estimated boundary.
1. A region of interest, ROI, segmentation system, comprising:
a display device configured to display an image;
a gaze-tracking device configured to generate gaze points relative to the displayed image;
a gaze point collector configured to select gaze points from the generated gaze points corresponding to a ROI of the displayed image;
a boundary identifier configured to estimate a boundary based on the selected gaze points; and
a region identifier configured to segment the ROI based on the generated boundary.
2. The ROI segmentation system according to claim 1 , wherein the gaze point collector is further configured to generate a dwelling point from the generated gaze points, and the dwelling point identifies a spatial location in the displayed image within the ROI, and the dwelling point defines a spatial and temporal proximity of the selected gaze points; and
wherein the display device is further configured to display a visual representation of the dwelling point relative to the displayed image.
3. The ROI segmentation system according to claim 1 , further including:
a computer input device which inputs a dwelling point identifying a spatial location in the displayed image within the ROI, and the dwelling point defines a temporal proximity of the selected gaze points; and
wherein the display device is further configured to display a visual representation of the dwelling point relative to the displayed image.
4. The ROI segmentation system according to claim 1 , further including:
a gaze point filter configured to filter the selected gaze points according to image characteristics of the dwelling point; and
wherein the boundary identifier is further configured to generate the boundary based on the filtered gaze points.
5. The ROI segmentation system according to claim 2 , wherein the gaze point filter filters using a maximum likelihood analysis including at least one of:
intensity similarity between a pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points;
gradient similarity between the pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points; or
a spatial distance from the pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points.
6. The ROI segmentation system according to claim 2 , wherein the display device is configured to display visual representations of the selected gaze points; and
the gaze point collector is further configured to calibrate a spatial location of the selected gaze points relative to the ROI based on an input from a computer input device identifying a location in the displayed image.
7. The ROI segmentation system according to claim 1 , wherein the boundary identifier is configured to generate the estimated boundary using at least one of an active contour algorithm and a level set algorithm.
8. The ROI segmentation system according to claim 1 , wherein the region identifier is configured to segment the ROI further based on at least one of:
a region growing algorithm;
an active contour algorithm;
a live wire algorithm; or
a level set algorithm.
9. The ROI segmentation system according to claim 1 , wherein the gaze-tracking device includes at least one of a stationary mounted device and a body worn device.
10. The ROI segmentation system according to claim 1 , wherein the displayed image is received from at least one of:
a picture archiving and communication system (PACS); and
a medical imaging device.
11. A method of region of interest, ROI, segmentation, comprising:
displaying an image;
generating gaze points relative to the displayed image;
selecting gaze points from the generated gaze points corresponding to a ROI of the displayed image;
estimating a boundary based on the selected gaze points; and
segmenting the ROI based on the estimated boundary.
12. The method of ROI segmentation according to claim 11 , wherein selecting includes:
inputting a dwelling point which identifies a spatial location in the displayed image within the ROI, and the dwelling point determined a temporal proximity of the selected gaze points; and
displaying a visual representation of the dwelling point relative to the image.
13. The method of ROI segmentation according to claim 11 , further including:
filtering the selected gaze points according to image characteristics of the dwelling point; and
wherein estimating the boundary is based on the filtered gaze points.
14. The method of ROI segmentation according to claim 11 , wherein filtering uses a maximum likelihood analysis including at least one of:
intensity similarity between a pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points;
gradient similarity between the pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points; or
a spatial distance from the pixel of the image corresponding to the dwelling point and pixels of the image corresponding to the selected gaze points.
15. The method of ROI segmentation according to claim 11 , further including:
displaying visual representations of the selected gaze points; and
calibrating a spatial location of the selected gaze points relative to the ROI based on an image location input from a computer input device.