US7020329B2
Expired / Lapsed
US
⏱️ 0.0 years remaining
Color image segmentation in an object recognition system
Assignee: Massachusetts Institute of Technology
•
Filing: 2002-08-30
•
Granted: 2006-03-28
•
Expires: 2022-08-30
Abstract
Abstract A method for segmenting a color image into a plurality of regions includes converting the image to a three-dimensional color space, generating a histogram for each dimension in the color space, using the histogram for each dimension to generate a plurality of connecting-boxes in the three-dimensional color space, and computing a normalized variance value of each connecting-box in order to form clusters of connecting-boxes corresponding to regions of the image. An illustrated application for the image segmentation method is in an object recognition system suitable for aligning a probe of an atomic force microscope with a feature of a sample.
Inventors
- 👤 Pradya Prempraneerach
- 👤 Kamal Youcef-Toumi
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 09:53
CPC Classifications
G; G06; G06T; G06T7/00; G06T7/10; G06T7/11; G06F; G06F18/00; G06F18/20; G06F18/23; G06F18/232; G06F18/2321; G06T7/136; G06T7/143; G06T7/155; G06T7/90; G06V; G06V10/00; G06V10/40; G06V10/56; G06T2207/00; G06T2207/10; G06T2207/10016
Claims Analysis
Total Claims: 20
Claim 1
Independent Claim
79 words
A method for segmenting a color image into a plurality of regions, comprising: converting the image to a three-dimensional color space; generating a histogram for each dimension in the color space; using the histogram for each dimension to generate a plurality of connecting-boxes in the three-dimensional color space; and computing a normalized variance value of each connecting-box in order to form clusters of connecting-boxes, each cluster corresponding to a respective region of the image in the three-dimensional color space.
Claim 2
Independent Claim
79 words
A method for segmenting a color image into a plurality of regions, comprising: converting the image to a three-dimensional color space; generating a histogram for each dimension in the color space; using the histogram for each dimension to generate a plurality of connecting-boxes in the three-dimensional color space; and computing a normalized variance value of each connecting-box in order to form clusters of connecting-boxes, each cluster corresponding to a respective region of the image in the three-dimensional color space.
Claim 3
Dependent Claim
12 words
The method of claim 1 wherein the histogram is a one-dimensional histogram.
Claim 4
Dependent Claim
17 words
The method of claim 1 wherein at least two of the plurality of connecting-boxes have different sizes.
Claim 5
Dependent Claim
38 words
The method of claim 1 wherein the computing step includes combining each of the plurality of connecting-boxes with a neighboring connecting-box on a path that has a local minimum normalized variance value in a steepest gradient descent direction.
Claim 6
Dependent Claim
29 words
The method of claim 1 wherein the histogram using step includes detecting a valley in each histogram, wherein each connecting-box is associated with a respective valley in the histogram.
Claim 7
Dependent Claim
19 words
The method of claim 5 wherein the valley detecting step comprises filtering the histogram to remove high frequency noise.
Claim 8
Dependent Claim
15 words
The method of claim 1 wherein the color space is a perceptually uniform color space.
Claim 9
Dependent Claim
21 words
The method of claim 7 wherein the perceptually uniform color space is an intensity hue saturation (I H S) color space.
Claim 10
Dependent Claim
18 words
The method of claim 1 further comprising the step of filtering the color image with an edge-preserving filter.
Claim 11
Dependent Claim
20 words
The method of claim 1 further comprising the step of filtering each of the histograms image with a morphological filter.
Claim 12
Dependent Claim
15 words
The method of claim 10 wherein the morphological filter comprises a raised triangular structuring element.
Claim 13
Dependent Claim
67 words
The method of claim 1 wherein forming clusters of connecting-boxes further comprises: grouping the plurality of connecting-boxes by determining a plurality of root nodes corresponding to the plurality of connecting-boxes having a local minimum normalized variance; and determining the branch direction from each connecting-box to a corresponding one of the plurality of root nodes by determining the steepest gradient direction of the minimum of the normalized variance.
Claim 14
Dependent Claim
20 words
The method of claim 1 further comprising labeling each pixel in said image to correspond to one of the clusters.
Claim 15
Dependent Claim
48 words
The method of claim 13 further comprising classifying said clusters to identify objects in said image, said classifying step comprising: training a classifier with images of a plurality of objects to be recognized; and recognizing ones of the plurality of objects to be recognized using the labeled pixels.
Claim 16
Dependent Claim
45 words
The method of claim 1 wherein the color image is of a probe and a sample, and said method further comprises: classifying the clusters of connecting-boxes into objects comprising the probe and the sample; and detecting a position of the probe relative to the sample.
Claim 17
Dependent Claim
104 words
An object recognition system for segmenting a color image into a plurality of regions, said system comprising: a color space converter operative to convert the image to a three-dimensional color space; a histogram generator operative to generate a histogram for each dimension in said three-dimensional color space; a histogram processor responsive to the histogram for each dimension to generate a plurality of connecting-boxes in the three-dimensional color space; and a clustering processor operative to compute the normalized variance value of each connecting-box in order to form clusters of connecting-boxes, each cluster corresponding to a respective region of the image in the three-dimensional color space.
Claim 18
Dependent Claim
104 words
An object recognition system for segmenting a color image into a plurality of regions, said system comprising: a color space converter operative to convert the image to a three-dimensional color space; a histogram generator operative to generate a histogram for each dimension in said three-dimensional color space; a histogram processor responsive to the histogram for each dimension to generate a plurality of connecting-boxes in the three-dimensional color space; and a clustering processor operative to compute the normalized variance value of each connecting-box in order to form clusters of connecting-boxes, each cluster corresponding to a respective region of the image in the three-dimensional color space.
Claim 19
Dependent Claim
27 words
The object recognition system of claim 16 further comprising a classifier responsive to said clusters of connecting-boxes for classifying said clusters to recognize objects in said image.
Claim 20
Dependent Claim
19 words
The object recognition system of claim 16 wherein said objects in said image comprise a probe and a sample.
References Cited (Backward Citations)
US4731863A
Cited
US5432893A
Cited
US6182069B1
Cited
US6009200A
Cited
US5933524A
Cited
US5848185A
Cited
US5764824A
Cited
US5825909A
Cited
US6108446A
Cited
US6178260B1
Cited
US6760488B1
Cited
US6788809B1
Cited
US20030044061A1
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2009-03-27 | 2009-09-26 | 2010-03-28 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2013-03-27 | 2013-09-26 | 2014-03-28 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2017-03-27 | 2017-09-26 | 2018-03-28 | $7700.00 | Window Closed / Lapsed |