Knowledge-based hierarchical method for detecting regions of interest

1. An automated method for detecting a region of interest in an image under consideration containing objects, comprising the steps of:

dividing prototypes into at least two groups based on prior knowledge;

generating at least one down-sampled copy of the image under consideration, each said at least one down-sampled copy of the image under consideration having a resolution lower than the resolution of the image under consideration;

automatically determining at which of said resolutions to detect objects corresponding to prototypes belonging to specific ones of said groups;

detecting at a first of said resolutions at least a first region containing at least one object from at least one of said groups;

detecting at a second of said resolutions at least a second region containing at least one object from another of said groups;

consolidating the plurality of regions; and

expanding the consolidated region to depict at least one region of interest.

2. The method of claim 1 wherein the expanding step comprises depicting one object in each region of interest.

3. The method of claim 1 further comprising the step of using a feature of said objects to aid in the detecting step.

4. The method of claim 1 wherein the detecting step, the consolidating step, and the expanding step each comprise the step of using prior knowledge.

5. The method of claim 1 wherein the detecting step further comprises the steps of:

splitting a region in the plurality of regions; and

merging at least two regions in the plurality of regions.

6. The method of claim 1 wherein the consolidating step comprises the step of using an arbitration process.

7. The method of claim 1 wherein the detecting step comprises the step of performing the detection of an object in a copy of the image under consideration in parallel with the detection of another object in a copy of the image under consideration.

8. The method of claim 1 further comprising the step of using prior knowledge to aid in said detecting step.

9. The method of claim 1 further comprising the step of performing probability assignments on each region of interest to modify the region of interest to conform to the shape of the object.

10. The method of claim 1 further comprising the step of designing a detection algorithm for each group of objects to be used in said detecting step.

11. The method of claim 1 wherein a first down-sampled copy of said image under consideration has a resolution different from a second down-sampled copy of said image under consideration.

12. The method of claim 11 further comprising the step of detecting at a third of said resolutions at least a third region containing at least one object from yet another of said groups.

13. A method for detecting a region of interest in an image under consideration, said method comprising the steps of:

generating at least one down-sampled copy of said image under consideration, each said at least one down-sampled copy of said image under consideration having a resolution different from the resolution of said image under consideration;

defining a plurality of prototypes, each of said prototypes characterized by at least one characteristic associated with said prototype;

determining a preferred one of said image under consideration and said at least one down-sampled copy of said image under consideration in which to detect potential objects of interest corresponding to specific ones of said prototypes;

detecting potential objects of interest in the corresponding preferred one(s) of said image under consideration and said at least one down-sampled copy of said image under consideration;

defining in each of said image under consideration and said at least one down-sampled copy of said image under consideration in which a potential object of interest was detected a region of candidacy corresponding to each of said potential objects of interest;

consolidating said regions of candidacy into said image under consideration; and

defining a region of interest about each of said regions of candidacy.

14. The method of claim 13 further comprising the step of removing unnecessary objects of interest in each down-sampled copy of said image under consideration before performing said step of combining.

15. The method of claim 13 wherein the step of determining is based on said characteristics of said prototypes and information about the sensor that was used to obtain the image under consideration.

16. The method of claim 13 wherein the step of generating comprises the step of using a wavelet transform.

17. The method of claim 13 wherein the step of detecting comprises the step of identifying the potential object of interest by its edge.

18. The method of claim 13 further comprising the step of splitting at least two potential objects of interest after said detecting step.

19. The method of claim 18 further comprising the step of merging a potential object of interest after said splitting step.

20. The method of claim 19 further comprising the step of performing probability assignments on each region of interest to modify the region of interest to conform to the shape of at least one potential object of interest.

21. The method of claim 13 wherein said step of consolidating comprises the step of using a multi-voting algorithm.

22. The method of claim 13 further comprising the step of expanding the region of interest.

23. The method of claim 13 further comprising the step of dividing said prototypes into groups based on said characteristics associated with each of said prototypes and designing an algorithm for detecting potential objects of interest corresponding to each of said groups.

24. The method of claim 13 wherein said step of determining is performed automatically.

25. The method of claim 13 wherein a first down-sampled copy of said image under consideration has a resolution different from a second down-sampled copy of said image under consideration.