Method of creating a divergence transform for identifying a feature of interest in hyperspectral data
View Patent ↗A system and method for identifying objects of interest in image data is provided. The present invention utilizes principles of Iterative Transformational Divergence in which objects in images, when subjected to special transformations, will exhibit radically different responses based on the physical, chemical, or numerical properties of the object or its representation (such as images), combined with machine learning capabilities. Using the system and methods of the present invention, certain objects that appear indistinguishable from other objects to the eye or computer recognition systems, or are otherwise almost identical, generate radically different and statistically significant differences in the image describers (metrics) that can be easily measured.
1. A method of creating a divergence transform for identifying a feature of interest in hyperspectral data, comprising:
(1) selecting a point operation;
(2) performing the point operation on a subset of the hyperspectral data that contains a known feature that is substantially similar to said feature of interest;
(3) adjusting the point operation if the point operation does not bifurcate the known feature;
(4) performing the adjusted point operation on the subset of the hyperspectral data; and
(5) repeating steps (3) and (4) until a resulting point operation bifurcates the known feature, whereby the resulting point operation corresponds to the divergence transform.
2. The method of claim 1 , wherein the point operation is at least partially a nonlinear point operation.
3. The method of claim 1 , wherein the point operation comprises at least one nodal point.
4. The method of claim 3 , wherein the at least one nodal point is set so as to effect the bifurcation of the feature of interest.
5. The method of claim 3 , wherein the nodal point is set so as to reduce background clutter in at least some of the hyperspectral data in the subset of hyperspectral data.
6. The method of claim 1 , wherein the resulting point operation is adapted to maintain an integrity of the feature of interest when the resulting point operation bifurcates the feature of interest.
7. The method of claim 1 , wherein the hyperspectral data comprises nonparametric hyperspectral data.
8. The method of claim 1 , wherein the hyperspectral data comprises two-dimensional hyperspectral data.
9. The method of claim 1 , wherein the hyperspectral data comprises three-dimensional hyperspectral data.
10. The method of claim 1 , wherein the hyperspectral data comprises x-ray hyperspectral data.
11. A method of creating a series of divergence transforms for identifying a feature of interest in hyperspectral data, comprising:
(1) selecting a series of point operations;
(2) performing the series of point operations sequentially on a subset of the hyperspectral data that contains a known feature that is substantially similar to said feature of interest;
(3) adjusting the series of point operations if the series of point operations do not bifurcate the known feature;
(4) performing the adjusted series of point operations on the subset of the hyperspectral data; and
(5) repeating steps (3) and (4) until a resulting series of point operations bifurcates the known feature, whereby the resulting series of point operations correspond to the series of divergence transforms.
12. The method of claim 11 , wherein the series of point operations are at least partially non-linear point operations.
13. The method of claim 11 , wherein each of the point operations in the series of point operations comprises at least one nodal point.
14. The method of claim 13 , wherein the nodal points are set so as to effect the bifurcation of the feature of interest.
15. The method of claim 13 , wherein the nodal points are set so as to reduce background clutter in at least some of the hyperspectral data in the subset of hyperspectral data.
16. The method of claim 11 , wherein the resulting series of point operations are adapted to maintain an integrity of the feature of interest when the resulting series of point operations bifurcates the feature of interest.
17. The method of claim 11 , wherein the hyperspectral data comprises nonparametric hyperspectral data.
18. The method of claim 11 , wherein the hyperspectral data comprises two-dimensional hyperspectral data.
19. The method of claim 11 , wherein the hyperspectral data comprises three-dimensional hyperspectral data.
20. The method of claim 11 , wherein the hyperspectral data comprises x-ray hyperspectral data.
21. A method of creating a divergence transform for identifying a feature of interest in hyperspectral data, comprising:
(1) assimilating a database of hyperspectral data, wherein at least some of the hyperspectral data contains a known feature that is substantially similar to said feature of interest;
(2) selecting a point operation;
(3) performing the point operation on a subset of the hyperspectral data that contains a known feature that is substantially similar to said feature of interest;
(4)adjusting the point operation if the point operation does not bifurcate the known feature;
(5) performing the adjusted point operation on the subset of the hyperspectral data; and
(6) repeating steps (4) and (5) until a resulting point operation bifurcates the known feature, whereby the resulting point operation corresponds to the divergence transform.
22. The method of claim 21 , wherein the point operation is at least partially a nonlinear point operation.
23. The method of claim 21 , wherein the point operation comprises at least one nodal point.
24. The method of claim 23 , wherein the at least one nodal point is set so as to effect the bifurcation of the feature of interest.
25. The method of claim 23 , wherein the nodal point is set so as to reduce background clutter in at least some of the hyperspectral data in the subset of hyperspectral data.
26. The method of claim 21 , wherein the resulting point operation is adapted to maintain an integrity of the feature of interest when the resulting point operation bifurcates the feature of interest.
27. The method of claim 21 , wherein the hyperspectral data comprises nonparametric hyperspectral data.
28. The method of claim 21 , wherein the hyperspectral data comprises two-dimensional hyperspectral data.
29. The method of claim 21 , wherein the hyperspectral data comprises three-dimensional hyperspectral data.
30. The method of claim 21 , wherein the hyperspectral data comprises x-ray hyperspectral data.
31. A method of creating a series of divergence transforms for identifying a feature of interest in hyperspectral data, comprising:
(1) assimilating a database of hyperspectral data, wherein at least some of the hyperspectral data contains a known feature that is substantially similar to said feature of interest;
(2) selecting a series of point operations;
(3) performing the series of point operations sequentially on a subset of the hyperspectral data that contains a known feature that is substantially similar to said feature of interest;
(4) adjusting the series of point operations if the series of point operations do not bifurcate the known feature;
(5) performing the adjusted series of point operations on the subset of the hyperspectral data; and
(6) repeating steps (4) and (5) until a resulting series of point operations bifurcates the known feature, whereby the resulting series of point operations correspond to the series of divergence transforms.
32. The method of claim 31 , wherein the series of point operations are at least partially non-linear point operations.
33. The method of claim 31 , wherein each of the point operations in the series of point operations comprises at least one nodal point.
34. The method of claim 33 , wherein the nodal points are set so as to effect the bifurcation of the feature of interest.
35. The method of claim 33 , wherein the nodal points are set so as to reduce background clutter in at least some of the hyperspectral data in the subset of hyperspectral data.
36. The method of claim 31 , wherein the resulting series of point operations are adapted to maintain an integrity of the feature of interest when the resulting series of point operations bifurcates the feature of interest.
37. The method of claim 31 , wherein the hyperspectral data comprises nonparametric hyperspectral data.
38. The method of claim 31 , wherein the hyperspectral data comprises two-dimensional hyperspectral data.
39. The method of claim 31 , wherein the hyperspectral data comprises three-dimensional hyperspectral data.
40. The method of claim 31 , wherein the hyperspectral data comprises x-ray hyperspectral data.