Automatic image segmentation methods and analysis
The invention provides methods and apparatus for image processing that perform image segmentation on data sets in two- and/or three-dimensions so as to resolve structures that have the same or similar grey values (and that would otherwise render with the same or similar intensity values) and that, thereby, facilitate visualization and processing of those data sets.
1. A method of processing anatomical structures present in a two-dimensional (2D) image slice using a digital data processor that labels connected components comprising:
(a) selecting a 2D image slice from a plurality of first 2D image slices spanning a three dimensional (3D) volume, using a digital data processing system configured for image processing, where each of the pixels of the selected 2D image slice have one or more of a first intensity value, a second intensity value and a third intensity value, where the second intensity value is greater than the first intensity value, where the third intensity value is greater than the second intensity value;
(b) computing with the digital data processing system a histogram of the 2D image slice;
(c) comparing the histogram with an anatomic atlas of average histograms including computing with the digital data processing system a correlation between the histogram and the anatomic atlas of average histograms;
(d) identifying one or more pixels corresponding with at least two structures of interest in the 2D image slice using the digital data processing system based on the correlation;
(e) assigning with the digital data processing system the one or more pixels identified in step (d) to one or more seed points, where the one or more pixels have an intensity value above the first intensity value or above the second intensity value, where a first seed point corresponds with a first structure of interest, where a second seed point corresponds with a second structure of interest;
(f) growing a region with the digital data processing system from the first seed point to at least one pixel connected to the first seed point until the at least one pixel overruns the second structure of interest to identify the first structure of interest; and
(g) labeling with the digital data processing system the first structure of interest on the selected 2D image slice.
2. The method of claim 1 , where in step (f) the region comprises gradually increasing an intensity of the first seed point.
3. The method of claim 1 , where the region continues to grow where the at least one pixel connected to the first seed point has an intensity value above the first intensity value.
4. The method of claim 3 , where the first structure of interest is resolved from a pixel not assigned to the first structure of interest based on assigning pixels at or above the first intensity value.
5. The method of claim 1 , where the first structure of interest is resolved from a pixel not assigned to the first structure of interest identified based on assigning pixels at or above the second intensity value.
6. The method of claim 1 , where step (c) includes the step of computing a cross correlation between the histogram and the anatomic atlas of average histograms.
7. The method of claim 1 , where the first structure of interest is selected from the group consisting of a bone structure and a vessel structure.
8. The method of claim 1 , where the first structure of interest corresponds to a bone structure is removed from the labeled 2D image slice.
9. The method of claim 1 , where the first structure of interest corresponds to a vessel structure is displayed in the labeled 2D image slice.
10. The method of claim 1 , further comprising where the at least one pixel not assigned to the first structure of interest are automatically removed from the labeled 2D image.
11. The method of claim 1 , further comprising where the at least one pixel not assigned to the first structure of interest are automatically removed from the labeled 2D image.
12. A method of processing anatomical structures present in a two-dimensional (2D) image slice using a digital data processor that labels connected components comprising:
(a) selecting a 2D image slice from a plurality of first 2D image slices spanning a three dimensional (3D) volume, using a digital data processing system configured for image processing, where each of the pixels of the selected 2D image slice have one or more of a first intensity value, a second intensity value and a third intensity value, where the second intensity value is greater than the first intensity value, where the third intensity value is greater than the second intensity value;
(b) computing with the digital data processing system a histogram of the 2D image slice;
(c) comparing the histogram with an anatomic atlas of average histograms including computing with the digital data processing system a correlation between the histogram and the anatomic atlas of average histograms;
(d) identifying one or more pixels corresponding with at least two structures of interest in the 2D image slice using the digital data processing system based on the correlation;
(e) assigning with the digital data processing system the one or more pixels identified in step (d) to one or more seed points, where the one or more pixels have an intensity value above the first intensity value or above the second intensity value, where a first seed point corresponds with a first structure of interest, where a second seed point corresponds with a second structure of interest;
(f) growing a region with the digital data processing system from the first seed point to at least one pixel connected to the first seed point until the at least one pixel overruns the second structure of interest to identify the first structure of interest; and
(g) coloring with the digital data processing system the first structure of interest on the selected 2D image slice.
13. The method of claim 12 , where in step (f) the region comprises gradually increasing an intensity of the first seed point.
14. The method of claim 12 , where the region continues to grow where the at least one pixel connected to the first seed point has an intensity value above the first intensity value.
15. The method of claim 14 , where the first structure of interest is resolved from a pixel not assigned to the first structure of interest based on assigning pixels at or above the first intensity value.
16. The method of claim 12 , where the first structure of interest is resolved from a pixel not assigned to the first structure of interest identified based on assigning pixels at or above the second intensity value.
17. The method of claim 12 , where step (c) includes the step of computing a cross correlation between the histogram and the anatomic atlas of average histograms.
18. The method of claim 12 , where the first structure of interest is selected from the group consisting of a bone structure and a vessel structure.
19. The method of claim 12 , where the first structure of interest corresponds to a bone structure is removed from the labeled 2D image slice.
20. The method of claim 12 , where the first structure of interest corresponds to a vessel structure is displayed in the labeled 2D image slice.