US5799100A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Computer-assisted method and apparatus for analysis of x-ray images using wavelet transforms
Assignee: University of South Florida St. Petersburg
•
Filing: 1996-06-03
•
Granted: 1998-08-25
•
Expires: 2016-06-03
Abstract
Abstract A computer-assisted diagnostic (CAD) method and apparatus are described for the enhancement and detection of suspicious regions in digital X-ray images, with particular emphasis on early cancer detection using digital mammography. An objective is to improve the sensitivity of detection of suspicious areas such as masses, while maintaining a low false positive detection rate, and to classify masses as benign or malignant. A modular CAD technique has been developed as a potentially automatic and/or second-opinion method for mass detection and classification in digital mammography that may in turn be readily modified for application with different digital X-ray detectors with varying gray-scale and resolution characteristics. The method consists of using a plurality of CAD modules to preprocess and enhance image features in the gray-level, the directional texture, and the morphological domains.
Inventors
- 👤 Laurence P. Clarke
- 👤 Wei Qian
- 👤 Lihua Li
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-25 10:09
CPC Classifications
G; G06; G06T; G06T7/00; G06T7/0002; G06T7/0012; Y; Y10; Y10S; Y10S128/00; Y10S128/92; Y10S128/922
Claims Analysis
Total Claims: 72
Claim 1
Independent Claim
97 words
A method for detecting and imaging a suspicious mass in a digital x-ray image, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; discriminating and extracting out from the directional features a spiculation emanating from the suspicious mass; and combining the spiculation with the suspicious area to produce a segmented image of the suspicious mass.
Claim 2
Independent Claim
97 words
A method for detecting and imaging a suspicious mass in a digital x-ray image, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; discriminating and extracting out from the directional features a spiculation emanating from the suspicious mass; and combining the spiculation with the suspicious area to produce a segmented image of the suspicious mass.
Claim 3
Dependent Claim
34 words
The method recited in claim 1, further comprising the step of suppressing image noise to produce a filtered image prior to the decomposing step, and wherein the decomposing step comprises decomposing the filtered image.
Claim 4
Dependent Claim
15 words
The method recited in claim 2, wherein the noise-suppressing step comprises applying a tree-structured filter.
Claim 5
Dependent Claim
23 words
The method recited in claim 3, wherein the tree-structured filter comprises three stages having a central weighted median filter as a subfiltering block.
Claim 6
Dependent Claim
16 words
The method recited in claim 2, wherein the noise-suppressing step comprises applying an adaptive multistage filter.
Claim 7
Dependent Claim
19 words
The method recited in claim 5, wherein the adaptive multistage filter comprises a two-stage filter incorporating an adaptive step.
Claim 8
Dependent Claim
26 words
The method recited in claim 6, wherein a first stage of the two-stage filter comprises a plurality of first-stage filters, each having a different window size.
Claim 9
Dependent Claim
32 words
The method recited in claim 7, wherein the first-stage filters comprise filters selected from the group consisting of: a linear filter, a tree-structured filter, an α-trimmed mean filter, and a median filter.
Claim 10
Dependent Claim
23 words
The method recited in claim 7, wherein the noise-suppressing step further comprises selecting an appropriate first-stage filter based upon a local signal activity.
Claim 11
Dependent Claim
17 words
The method recited in claim 1, wherein the decomposing step comprises using a multiorientation directional wavelet transform.
Claim 12
Dependent Claim
36 words
The method recited in claim 10, wherein the decomposing step comprises using a multiorientation directional wavelet transform to decompose the image in the Fourier domain of support into a plurality of wavelets having differing orientation selectivities.
Claim 13
Dependent Claim
33 words
The method recited in claim 11, wherein the image decomposing step comprises decomposing the image in the Fourier domain of support into a number of wavelets comprising an integral multiple of eight wavelets.
Claim 14
Dependent Claim
32 words
The method recited in claim 10, wherein the decomposing step comprises using an N-directional wavelet transform for directionally enhancing the image, wherein N is chosen to match a resolution of the image.
Claim 15
Dependent Claim
35 words
The method recited in claim 10, wherein the decomposing step comprises using an N-directional wavelet transform for feature extraction from the image, wherein N is chosen to extract a desired feature resolution of the image.
Claim 16
Dependent Claim
33 words
The method recited in claim 1, further comprising the step, prior to the selectively enhancing step, of segmenting the smoothed image into a first region comprising tissue and a second region comprising background.
Claim 17
Dependent Claim
20 words
The method recited in claim 1, wherein the selectively enhancing step comprises performing selective image enhancement on the smoothed image.
Claim 18
Dependent Claim
20 words
The method recited in claim 16, wherein the selective image enhancement performing step comprises using an M-channel multiresolution wavelet transform.
Claim 19
Dependent Claim
29 words
The method recited in claim 17, wherein the multiresolution wavelet transform using step comprises using an M-channel tree-structured wavelet transform, wherein M is greater than or equal to 2.
Claim 20
Dependent Claim
26 words
The method recited in claim 18, wherein the M-channel tree-structured wavelet transform using step comprises implementing the M-channel tree-structured wavelet transform on quadrature mirror filter banks.
Claim 21
Dependent Claim
41 words
The method recited in claim 18, wherein the M-channel tree-structured wavelet transform using step comprises implementing the M-channel tree-structured wavelet transform on quadrature mirror filter banks, wherein M is an even integer, for improving subsequent image enhancement and suspicious area segmentation.
Claim 22
Dependent Claim
18 words
The method recited in claim 17, wherein the segmenting step comprises performing adaptive clustering on the enhanced image.
Claim 23
Dependent Claim
31 words
The method recited in claim 21, wherein the adaptive clustering performing step comprises maximizing a preservation of image details of the segmented suspicious area, for improving the retention of shape-related features.
Claim 24
Dependent Claim
51 words
The method recited in claim 1, wherein the discriminating and extracting step comprises applying a tracing algorithm to the segmented suspicious area, the tracing algorithm comprising a contour-tracing loop for locating spicules intersecting the suspicious area and a spicule-tracing loop for determining a length and a number of the located spicules.
Claim 25
Dependent Claim
61 words
The method recited in claim 23, wherein the contour-tracing loop applying step comprises: tracing a contour of the suspicious area; detecting a directional feature having an intersection with the suspicious area contour; and comparing a feature intensity of the intersecting directional feature with a predetermined intensity range, a feature intensity falling within the intensity range indicating the presence of a spicule.
Claim 26
Independent Claim
89 words
A method for detecting and classifying a suspicious area in a digital x-ray image, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and classifying the suspicious area from the feature analysis.
Claim 27
Independent Claim
89 words
A method for detecting and classifying a suspicious area in a digital x-ray image, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and classifying the suspicious area from the feature analysis.
Claim 28
Dependent Claim
32 words
The method recited in claim 25, wherein the feature analysis and classification steps comprise analyzing a shape of the suspicious area and classifying the suspicious area as round, oval, lobulated, or irregular.
Claim 29
Dependent Claim
35 words
The method recited in claim 26, wherein the feature analysis and classification steps comprise analyzing a mass margin of the suspicious area for fuzziness and classifying a probability of malignancy based upon the fuzziness analysis.
Claim 30
Dependent Claim
35 words
The method recited in claim 25, wherein the feature analysis and classification steps comprise analyzing a mass density criterion of the suspicious area, the mass density criterion providing an indication of a probability of malignancy.
Claim 31
Dependent Claim
21 words
The method recited in claim 28, wherein the mass density analysis step comprises calculating a density uniformity within the suspicious area.
Claim 32
Dependent Claim
48 words
The method recited in claim 28, wherein the mass density analysis step comprises calculating an average mass density within the suspicious area and comparing the average mass density with a mass density of tissue surrounding the suspicious area, the comparison providing an indication of a probability of malignancy.
Claim 33
Independent Claim
109 words
A method for detecting and classifying a suspicious area in a digital x-ray image to detect a suspicious mass, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and classifying the suspicious area from the feature analysis.
Claim 34
Independent Claim
109 words
A method for detecting and classifying a suspicious area in a digital x-ray image to detect a suspicious mass, the method comprising the steps of: decomposing the image using a directional wavelet transform; smoothing the decomposed image by removing directional features from the decomposed image; selectively enhancing a suspicious area in the smoothed image; segmenting the suspicious area in the smoothed image; detecting a suspicious mass in the suspicious area; discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and classifying the suspicious area from the feature analysis.
Claim 35
Dependent Claim
58 words
The method recited in claim 31, wherein: the directional feature discriminating and extracting step comprises discriminating and extracting out all detected spiculations emanating from the suspicious area; and the feature analysis and classification steps comprise analyzing a mass margin of the suspicious area for a spiculation criterion, for classifying a probability of malignancy based upon the spiculation criterion.
Claim 36
Dependent Claim
38 words
The method recited in claim 32, wherein the mass margin analysis step comprises analyzing at least one of a set of spiculation criteria chosen from a number of spiculations, an average speculation length, and a spiculation angular distribution.
Claim 37
Independent Claim
107 words
An apparatus for detecting a suspicious mass a digital x-ray image, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image comprising means for removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; and means for combining the spiculation with the suspicious area to produce a segmented image of a suspicious mass.
Claim 38
Independent Claim
107 words
An apparatus for detecting a suspicious mass a digital x-ray image, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image comprising means for removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; and means for combining the spiculation with the suspicious area to produce a segmented image of a suspicious mass.
Claim 39
Dependent Claim
35 words
The apparatus recited in claim 34, further comprising means for suppressing image noise to produce a filtered image prior to the decomposing step, and wherein the decomposing means comprises means for decomposing the filtered image.
Claim 40
Dependent Claim
14 words
The apparatus recited in claim 35, wherein the noise-suppressing means comprises a tree-structured filter.
Claim 41
Dependent Claim
23 words
The apparatus recited in claim 36, wherein the tree-structured filter comprises three stages having a central weighted median filter as a subfiltering block.
Claim 42
Dependent Claim
15 words
The apparatus recited in claim 35, wherein the noise-suppressing means comprises an adaptive multistage filter.
Claim 43
Dependent Claim
19 words
The apparatus recited in claim 38, wherein the adaptive multistage filter comprises a two-stage filter incorporating an adaptive algorithm.
Claim 44
Dependent Claim
26 words
The apparatus recited in claim 39, wherein a first stage of the two-stage filter comprises a plurality of first-stage filters, each having a different window size.
Claim 45
Dependent Claim
32 words
The apparatus recited in claim 40, wherein the first-stage filters comprise filters selected from the group consisting of: a linear filter, a tree-structured filter, an α-trimmed mean filter, and a median filter.
Claim 46
Dependent Claim
25 words
The apparatus recited in claim 40, wherein the noise-suppressing means further comprises means for selecting an appropriate first-stage filter based upon a local signal activity.
Claim 47
Dependent Claim
16 words
The apparatus recited in claim 36, wherein the decomposing means comprises a multiorientation directional wavelet transform.
Claim 48
Dependent Claim
33 words
The apparatus recited in claim 43, wherein the decomposing means comprises a multiorientation directional wavelet transform for decomposing the image in the Fourier domain of support into eight wavelets having differing orientation selectivities.
Claim 49
Dependent Claim
31 words
The apparatus recited in claim 43, wherein the decomposing means comprises an N-directional wavelet transform for directionally enhancing the image, wherein N is chosen to match a resolution of the image.
Claim 50
Dependent Claim
34 words
The apparatus recited in claim 43, wherein the decomposing means comprises an N-directional wavelet transform for feature extraction from the image, wherein N is chosen to extract a desired feature resolution of the image.
Claim 51
Dependent Claim
34 words
The apparatus recited in claim 43, wherein the decomposing means comprises means for decomposing the image in the Fourier domain of support into a number of wavelets comprising an integral multiple of eight wavelets.
Claim 52
Dependent Claim
26 words
The apparatus recited in claim 34, further comprising means for segmenting the smoothed image into a first region comprising tissue and a second region comprising background.
Claim 53
Dependent Claim
16 words
The apparatus recited in claim 34, wherein the selectively enhancing means comprises a multiresolution wavelet transform.
Claim 54
Dependent Claim
17 words
The apparatus recited in claim 49, wherein the multiresolution wavelet transform comprises an M-channel tree-structured wavelet transform.
Claim 55
Dependent Claim
26 words
The apparatus recited in claim 50, wherein the M-channel tree-structured wavelet transform comprises means for implementing the M-channel tree-structured wavelet transform on quadrature mirror filter banks.
Claim 56
Dependent Claim
20 words
The apparatus recited in claim 34, wherein the segmenting means comprises means for performing adaptive clustering on the enhanced-area image.
Claim 57
Dependent Claim
37 words
The apparatus recited in claim 52, wherein the adaptive clustering performing means comprises means for performing an initial segmentation followed by iteratively preserving image details of the segmented suspicious area, for improving the retention of shape-related features.
Claim 58
Dependent Claim
51 words
The apparatus recited in claim 34, wherein the discriminating and extracting means comprises tracing algorithm for application to the segmented suspicious area, the tracing algorithm comprising a contour-tracing loop for locating spicules intersecting the suspicious area and a spicule-tracing loop for determining a length and a number of the located spicules.
Claim 59
Dependent Claim
65 words
The apparatus recited in claim 54, wherein the contour-tracing loop comprises: means for tracing a contour of the suspicious area; means for detecting a directional feature having an intersection with the suspicious area contour; and means for comparing a feature intensity of the intersecting directional feature with a predetermined intensity range, a feature intensity falling within the intensity range indicating the presence of a spicule.
Claim 60
Independent Claim
100 words
An apparatus for detecting and classifying a suspicious area in a digital x-ray image, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image by removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and means for classifying the suspicious area from the feature analysis.
Claim 61
Independent Claim
100 words
An apparatus for detecting and classifying a suspicious area in a digital x-ray image, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image by removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and means for classifying the suspicious area from the feature analysis.
Claim 62
Dependent Claim
36 words
The apparatus recited in claim 56, wherein the feature analysis and classification means comprise means for analyzing a shape of the suspicious area and means for classifying the suspicious area as round, oval, lobulated, or irregular.
Claim 63
Dependent Claim
39 words
The apparatus recited in claim 57, wherein the feature analysis and classification means comprise means for analyzing a mass margin of the suspicious area for fuzziness and means for classifying a probability of malignancy based upon the fuzziness analysis.
Claim 64
Dependent Claim
37 words
The apparatus recited in claim 56, wherein the feature analysis and classification means comprise means for analyzing a mass density criterion of the suspicious area, the mass density criterion providing an indication of a probability of malignancy.
Claim 65
Dependent Claim
23 words
The apparatus recited in claim 59, wherein the mass density analysis means comprises means for calculating a density uniformity within the suspicious area.
Claim 66
Dependent Claim
52 words
The apparatus recited in claim 59, wherein the mass density analysis means comprises means for calculating an average mass density within the suspicious area and means for comparing the average mass density with a mass density of tissue surrounding the suspicious area, the comparison providing an indication of a probability of malignancy.
Claim 67
Independent Claim
122 words
An apparatus for detecting and classifying a suspicious area in a digital x-ray image to detect a suspicious mass, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image by removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for discriminating and extracting out from the directional features a speculation emanating from the suspicious area; means for analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and means for classifying the suspicious area from the feature analysis.
Claim 68
Independent Claim
122 words
An apparatus for detecting and classifying a suspicious area in a digital x-ray image to detect a suspicious mass, the apparatus comprising: means for decomposing the image comprising a directional wavelet transform; means for smoothing the decomposed image by removing directional features from the decomposed image; means for selectively enhancing a suspicious area in the smoothed image; means for segmenting the suspicious area in the smoothed image; means for detecting a suspicious mass in the suspicious area; means for discriminating and extracting out from the directional features a speculation emanating from the suspicious area; means for analyzing a feature of the segmented suspicious area having a correlation with a classification criterion; and means for classifying the suspicious area from the feature analysis.
Claim 69
Dependent Claim
62 words
The apparatus recited in claim 62, wherein: the directional feature discriminating and extracting means comprises means for discriminating and extracting out all detected spiculations emanating from the suspicious area; and the feature analysis and classification means comprise means for analyzing a mass margin of the suspicious area for a spiculation criterion, for classifying a probability of malignancy based upon the spiculation criterion.
Claim 70
Dependent Claim
40 words
The apparatus recited in claim 63, wherein the mass margin analysis means comprises means for analyzing at least one of a set of spiculation criteria chosen from a number of spiculations, an average spiculation length, and a speculation angular distribution.
Claim 71
Independent Claim
131 words
A method for preprocessing a raw x-ray mammographic image for detecting a suspicious mass, the method comprising the steps of: producing a gray-level domain image by suppressing image noise in the raw image; decomposing the gray-level domain image using a directional wavelet transform; smoothing the decomposed image by removing directional features therefrom; detecting and enhancing a suspicious area in the smoothed image; producing a morphological domain image by segmenting the suspicious area in the smoothed image; discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; and producing a directional texture domain image by analyzing the spiculation; wherein the images in the gray-level domain, the morphological domain, and the directional texture domain are useful as input to a processing module for assisting in suspicious mass detection.
Claim 72
Independent Claim
131 words
A method for preprocessing a raw x-ray mammographic image for detecting a suspicious mass, the method comprising the steps of: producing a gray-level domain image by suppressing image noise in the raw image; decomposing the gray-level domain image using a directional wavelet transform; smoothing the decomposed image by removing directional features therefrom; detecting and enhancing a suspicious area in the smoothed image; producing a morphological domain image by segmenting the suspicious area in the smoothed image; discriminating and extracting out from the directional features a spiculation emanating from the suspicious area; and producing a directional texture domain image by analyzing the spiculation; wherein the images in the gray-level domain, the morphological domain, and the directional texture domain are useful as input to a processing module for assisting in suspicious mass detection.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-08-24 | 2002-02-23 | 2002-08-25 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-08-24 | 2006-02-23 | 2006-08-25 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-08-24 | 2010-02-23 | 2010-08-25 | $7700.00 | Window Closed / Lapsed |