Hot spot detection, segmentation and identification in pet and spect images

1. An apparatus comprising:

a potential region of interest segmentation device which segments an image data into regions of potential interest, the potential region of interest segmentation device performing a watershed transform which segments the image data in a topographic fashion starting at seed points and propagating until two regions including two adjacent seed points meet;

an identifying device which, from the regions of potential interest, identifies a region of interest including an object of interest based on a set of rules; and

a recognizing device which differentiates among the identified objects of interest and selects at least one particular object of interest.

2. The apparatus as set forth in claim 1 , further including:

an object segmentation device which determines boundaries of the selected object of interest based on a localized threshold determined from the region of interest, which includes the selected object of interest, and isolates the selected object of interest from other tissue.

3. The apparatus as set forth in claim 1 , wherein the watershed transform starts at the seed points with highest grayscale values and propagates down circumferentially in grayscale values such that regions dividing lines are formed where pixels with lowest grayscale values are found.

4. The apparatus as set forth in claim 1 , wherein the set of rules includes at least one of:

rules for identifying a heart;

rules for identifying a tumor;

rules for identifying a liver; and

rules excluding artifacts.

5. The apparatus as set forth in claim 4 , further including:

a criteria selecting device which selects rules for identifying the region of interest based at least on an associated object of interest.

6. One of a PET and SPECT imaging system for performing the steps of (1) segmenting an image data into regions of potential interest, (2) from the regions of potential interest, identifying a region of interest including one or more objects of interest identified based on a set of rules, (3) differentiating among the identified objects of interest, and (4) selecting at least one particular object of interest.

7. The system as set forth in claim 6 , wherein the selecting of at least one object of interest includes:

determining first, second and third principal axes in the identified object of interest;

scanning along determined principal axes; and

counting a number of maxima points, the object of interest being selected based on the count.

8. The system as set forth in claim 7 , wherein, based on the count, the selected object of interest includes at least one of a left ventricle of a heart and a tumor.

9. The system as set forth in claim 7 , wherein the selecting further includes:

determining, based on the count, a long axis and short axes in the identified object of interest which short axes lie perpendicularly to the long axis.

10. The system as set forth in claim 6 , further performing the steps of:

determining boundaries of the selected object of interest based on localized threshold determined from the region of interest which includes the selected object of interest; and

isolating the selected object of interest from other tissue.

11. The system as set forth in claim 6 , wherein the set of rules includes at least one of:

rules for identifying a heart;

rules for identifying a tumor;

rules for identifying a liver; and

rules excluding artifacts.

12. The system as set forth in claim 11 , further performing the steps of:

selecting rules for identifying the region of interest based at least on a particular object of interest.

13. An imaging method comprising:

segmenting an image data into regions of potential interest, wherein the segmenting includes performing a watershed transform;

from the regions of potential interest, identifying a region of interest including one or more objects of interest identified based on a set of rules;

differentiating among the identified objects of interest; and

selecting at least one particular object of interest.

14. The method as set forth in claim 13 , further including:

starting at points with highest grayscale values;

propagating regions down in grayscale values; and

defining regions dividing lines.

15. An imaging method comprising:

segmenting an image data into regions of potential interest;

from the regions of potential interest, identifying a region of interest including one or more objects of interest identified based on a set of rules;

differentiating among the identified objects of interest wherein the differentiating includes:

determining principal axes of rotation in the identified object of interest,

scanning along determined principal axes,

counting a number of maxima points, and

selecting the object of interest based on the count; and

selecting at least one particular object of interest.

16. The method as set forth in claim 15 , wherein, based on the count, the selected object of interest includes at least one of a left ventricle of a heart and a tumor.

17. The method as set forth in claim 15 , further including:

based on the count, determining a longest axis in the identified object of interest; and

determining at least one short axis in the identified object of interest which short axis lies perpendicularly to the longest axis.

18. A processor programmed to perform the steps of:

segmenting an image data into regions of potential interest;

based on a set of rules, identifying a particular object of interest in one region of potential interest;

fitting a pre-selected model with the identified object of interest;

determining a long axis in the identified object of interest; and

determining two short axes in the identified object of interest, the short axes extending perpendicularly to the long axis.