Image processing using measures of similarity
First Claim
Patent Images
1. A method of locating a portion of tissue with a characteristic of interest, the method comprising the steps of:
- (a) characterizing an acetowhitening signal from a temporal sequence of images of a tissue following application of a chemical agent to the tissue, wherein the chemical agent comprises acetic acid;
(b) analyzing the acetowhitening signal to determine a measure of similarity between two selected regions of the tissue, the measure of similarity indicating how similarly tissue in each region responds to the chemical agent, and grouping the two selected regions if the measure of similarity is larger than a given threshold, wherein the step of determining the measure of similarity comprises, for each of the two selected regions, averaging data corresponding to pixels within the region at each of a plurality of time steps to obtain a mean signal for the region, then quantifying the similarity between the two resulting, mean signals;
(c) repeating step (b), thereby differentiating regions according to how tissue in each region responds to the chemical agent; and
(d) locating a portion of the tissue with a characteristic of interest, the located portion corresponding to at least one of the differentiated regions.
4 Assignments
0 Petitions
Accused Products
Abstract
The invention provides methods of relating a plurality of images based on measures of similarity. The methods of the invention are useful in the segmentation of a sequence of colposcopic images of tissue, for example. The methods may be applied in the determination of tissue characteristics in acetowhitening testing of cervical tissue, for example.
-
Citations
34 Claims
-
1. A method of locating a portion of tissue with a characteristic of interest, the method comprising the steps of:
-
(a) characterizing an acetowhitening signal from a temporal sequence of images of a tissue following application of a chemical agent to the tissue, wherein the chemical agent comprises acetic acid; (b) analyzing the acetowhitening signal to determine a measure of similarity between two selected regions of the tissue, the measure of similarity indicating how similarly tissue in each region responds to the chemical agent, and grouping the two selected regions if the measure of similarity is larger than a given threshold, wherein the step of determining the measure of similarity comprises, for each of the two selected regions, averaging data corresponding to pixels within the region at each of a plurality of time steps to obtain a mean signal for the region, then quantifying the similarity between the two resulting, mean signals; (c) repeating step (b), thereby differentiating regions according to how tissue in each region responds to the chemical agent; and (d) locating a portion of the tissue with a characteristic of interest, the located portion corresponding to at least one of the differentiated regions. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A method of differentiating regions of a tissue, the method comprising the steps of:
-
(a) accessing a temporal sequence of images of a tissue following application of a chemical agent to the tissue; and (b) creating a segmentation mask that represents an image plane divided into regions according to how similarly tissue in each region responds to the chemical agent, wherein step (b) comprises analyzing an acetowhitening signal to determine a measure of similarity between two selected regions of the tissue, the measure of similarity indicating how similarly tissue in each region responds to the chemical agent, and grouping the two selected regions if the measure of similarity is larger than a given threshold, wherein the step of determining the measure of similarity comprises, for each of the two selected regions, averaging data corresponding to pixels within the region at each of a plurality of time steps to obtain a mean signal for the region, then quantifying the similarity between the two resulting mean signals. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24, 25, 26)
-
-
27. A system for differentiating regions of a tissue, the system comprising:
-
a light source that illuminates a tissue; a camera that obtains a temporal sequence of images of the tissue following application of a chemical agent to the tissue, the chemical agent comprising acetic acid; and software that performs the steps of; (i) characterizing an acetowhitening signal from temporal sequence of images; (ii) analyzing the acetowhitening signal to determining a measure of similarity between two selected regions of the tissue, the measure of similarity indicating how similarly tissue in each region responds to the chemical agent, and grouping the two selected regions if the measure of similarity is larger than a given threshold, wherein the step of determining the measure of similarity comprises, for each of the two selected regions, averaging data corresponding to pixels within the region at each of a plurality of time steps to obtain a mean signal for the region, then quantifying the similarity between the two resulting mean signals; and (iii) repeating step (ii), thereby differentiating regions according to how tissue in each region responds to the chemical agent. - View Dependent Claims (28, 29, 30, 31, 32, 33, 34)
-
Specification