Characterizing a texture of an image
View Patent ↗Among other things, a texture of an image is characterized by deriving lacunarity measures from the image based on a wavelet analysis.
1. A machine-based method comprising:
analyzing wavelet maxima in a wavelet maxima representation of an image; and
characterizing a texture of the image by deriving lacunarity measures based on the analysis of the wavelet maxima.
2. The method of claim 1 in which the image comprises a multispectral image.
3. The method of claim 2 in which the image comprises an image of a biological tissue.
4. The method of claim 1 in which the image is a gray scale image and the method also including computing the wavelet maxima representation for the gray scale image.
5. The method of claim 1 in which the image comprises an analysis region having a skin lesion.
6. The method of claim 5 also including applying automatic segmentation to create a mask of the lesion.
7. The method of claim 1 in which the lacunarity measures are derived at least in part based on a gliding box method.
8. The method of claim 7 in which the gliding box method uses a window of fixed characteristic size R.
9. The method of claim 8 in which the window comprises a circular window.
10. The method of claim 8 in which analyzing the wavelet maxima comprises counting the wavelet maxima in the window to generate a distribution of the counts indexed by a wavelet level L.
11. The method of claim 10 in which the lacunarity measures are indexed by the wavelet level L and the window of fixed characteristic size R.
12. The method of claim 8 in which the lacunarity measures are determined as a function of R.
13. The method of claim 1 also including testing the lacunarity measures for separability classes in which images are grouped.
14. The method of claim 6 in which the lacunarity measures are derived at least in part based on a gliding box method using a window of fixed characteristic size R, the window being inside the mask without extending beyond borders of the mask.
15. The method of claim 8 in which analyzing the wavelet maxima comprises computing a probability distribution of a density of the wavelet maxima in the wavelet maxima representation of the image.
16. The method of claim 15 in which the lacunarity measures are derived based on a change in the probability distribution of the density of the wavelet maxima as a function of R.
17. The method of claim 15 also including computing additional probability distributions of the density of wavelet maxima in the wavelet maxima representation using windows having different characteristic sizes and normalizing all computed probability distributions.
18. The method of claim 10 also including estimating a local similarity in the image by plotting the wavelet maxima counts as a function of the wavelet level L on a logarithmic scale.
19. A non-transitory computer-readable medium having encoded thereon instructions for performing operations comprising:
analyzing wavelet maxima in a wavelet maxima representation of an image; and
characterizing a texture of the image by deriving lacunarity measures based on the analysis of the wavelet maxima.
20. The non-transitory computer-readable medium of claim 19 in which the image comprises a multispectral image.
21. The non-transitory computer-readable medium of claim 20 in which the image comprises an image of a biological tissue.
22. The non-transitory computer-readable medium of claim 19 in which the image is a gray scale image and operations also including computing the wavelet maxima representation for the gray scale image.
23. The non-transitory computer-readable medium of claim 19 in which the image comprises an analysis region having a skin lesion.
24. The non-transitory computer-readable medium of claim 23 in which the operations also include applying automatic segmentation to create a mask of the lesion.
25. The non-transitory computer-readable medium of claim 19 in which the lacunarity measures are derived at least in part based on a gliding box method.
26. The non-transitory computer-readable medium of claim 25 in which the gliding box method uses a window of fixed characteristic size R.
27. The non-transitory computer-readable medium of claim 26 in which the window comprises a circular window.
28. The non-transitory computer-readable medium of claim 26 in which analyzing the wavelet maxima comprises counting the wavelet maxima in the window to generate a distribution of the counts indexed by a wavelet level L.
29. The non-transitory computer-readable medium of claim 28 in which the lacunarity measures are indexed by the wavelet level L and the window of fixed characteristic size R.
30. The non-transitory computer-readable medium of claim 26 in which the lacunarity measures are determined as a function of R.
31. The non-transitory computer-readable medium of claim 19 in which the operations also include testing the lacunarity measures for separability classes in which the image is grouped.
32. The non-transitory computer-readable medium of claim 24 in which the lacunarity measures are derived at least in part based on a gliding box method using a window of fixed characteristic size R, the window being inside the mask without extending beyond borders of the mask.
33. The non-transitory computer-readable medium of claim 26 in which analyzing the wavelet maxima comprises computing a probability distribution of a density of the wavelet maxima in the wavelet maxima representation of the image.
34. The non-transitory computer-readable medium of claim 33 in which the lacunarity measures are derived based on a change in the probability distribution of the density of the wavelet maxima as a function of R.
35. The non-transitory computer-readable medium of claim 33 in which the operations also include computing additional probability distributions of the density of wavelet maxima in the wavelet maxima representation using windows having different characteristic sizes and normalizing all computed probability distributions.
36. The non-transitory computer-readable medium of claim 28 also including estimating a local similarity in the image by plotting the wavelet maxima counts as a function of the wavelet level L on a logarithmic scale.
37. A machine comprising:
a processor;
a storage device that stores a program for execution by the processor, the program comprising instructions for causing the processor to
analyze wavelet maxima in a wavelet maxima representation of an image; and
characterize a texture of the image by deriving lacunarity measures based on the analysis of the wavelet maxima.
38. The machine of claim 37 in which the image comprises a multispectral image.
39. The machine of claim 38 in which the image comprises an image of a biological tissue.
40. The machine of claim 37 in which the image is a gray scale image and the processor is also caused to compute the wavelet maxima representation for the gray scale image.
41. The machine of claim 37 in which the image comprises an analysis region having a skin lesion.
42. The machine of claim 41 in which the processor is also caused to apply automatic segmentation to create a mask of the lesion.
43. The machine of claim 37 in which the lacunarity measures are derived at least in part based on a gliding box method.
44. The machine of claim 43 in which the gliding box method uses a window of fixed characteristic size R.
45. The machine of claim 44 in which the window comprises a circular window.
46. The machine of claim 44 in which analyzing the wavelet maxima comprises counting the wavelet maxima in the window to generate a distribution of the counts indexed by a wavelet level L.
47. The machine of claim 46 in which the lacunarity measures are indexed by the wavelet level L and the window of fixed characteristic size R.
48. The machine of claim 44 in which the lacunarity measures are determined as a function of R.
49. The machine of claim 37 in which the processor is also caused to test the lacunarity measures for separability classes in which the image is grouped.
50. The machine of claim 42 in which the lacunarity measures are derived at least in part based on a gliding box method using a window of fixed characteristic size R, the window being inside the mask without extending beyond borders of the mask.
51. The machine of claim 44 in which analyzing the wavelet maxima comprises computing a probability distribution of a density of the wavelet maxima in the wavelet maxima representation of the image.
52. The machine of claim 51 in which the lacunarity measures are derived based on a change in the probability distribution of the density of the wavelet maxima as a function of R.
53. The machine of claim 51 in which the processor is also caused to compute additional probability distributions of the density of wavelet maxima in the wavelet maxima representation using windows having different characteristic sizes and to normalize all computed probability distributions.
54. The machine of claim 46 also including estimating a local similarity in the image by plotting the wavelet maxima counts as a function of the wavelet level L on a logarithmic scale.