Optimal, user-friendly, object background separation
First Claim
1. A method of identifying an object of interest in digital images, the method comprising:
- a. obtaining first samples of an intensity distribution of one or more objects of interest in one or more of the digital images based upon one or more wavelength bands;
b. obtaining second samples of an intensity distribution of confounder objects in one or more of the digital images, at a predetermined frequency;
c. transforming the first and second samples into an appropriate first space;
d. performing dimensionality factor reduction on the transformed first and second samples, whereby the dimensionality factor reduction of the transformed first and second samples generates an object detector;
e. transforming one or more of the digital images into the first space;
f. performing dimensionality factor reduction on the transformed digital images, whereby the dimensionality factor reduction of the transformed digital images generates one or more reduced images;
g. classifying one or more pixels of the one or more reduced images based on a comparison with the object detector, wherein the classification comprises;
locating one or more K samples that minimize a distance to pixels in a pre-defined neighborhood; and
classifying one or more pixels as one of abnormal and normal, using the distance to the K samples and a label associated with the K samples, wherein a lesion likelihood index Lp for a pixel p of neighborhood Np is obtained from the K nearest samples (Sk)k=1 . . . K, with labels (lk)k=1 . . . K equal to 1 for lesions and −
1 for negative lesion confounders, by the following formula Lp=Σ
k=1Klkexp(−
∥
Np−
Sk∥
) and used to automatically classify the pixel p as abnormal or normal;
andh. identifying one or more objects of interest in the reduced digital images from the classified pixels.
5 Assignments
0 Petitions
Accused Products
Abstract
A method of identifying an object of interest can comprise obtaining first samples of an intensity distribution of one or more object of interest, obtaining second samples of an intensity distribution of confounder objects, transforming the first and second samples into an appropriate first space, performing dimension reduction on the transformed first and second samples, whereby the dimension reduction of the transformed first and second samples generates an object detector, transforming one or more of the digital images into the first space, performing dimension reduction on the transformed digital images, whereby the dimension reduction of the transformed digital images generates one or more reduced images, classifying one or more pixels of the one or more reduced images based on a comparison with the object detector, and identifying one ore more objects of interest from the classified pixels.
94 Citations
27 Claims
-
1. A method of identifying an object of interest in digital images, the method comprising:
-
a. obtaining first samples of an intensity distribution of one or more objects of interest in one or more of the digital images based upon one or more wavelength bands; b. obtaining second samples of an intensity distribution of confounder objects in one or more of the digital images, at a predetermined frequency; c. transforming the first and second samples into an appropriate first space; d. performing dimensionality factor reduction on the transformed first and second samples, whereby the dimensionality factor reduction of the transformed first and second samples generates an object detector; e. transforming one or more of the digital images into the first space; f. performing dimensionality factor reduction on the transformed digital images, whereby the dimensionality factor reduction of the transformed digital images generates one or more reduced images; g. classifying one or more pixels of the one or more reduced images based on a comparison with the object detector, wherein the classification comprises; locating one or more K samples that minimize a distance to pixels in a pre-defined neighborhood; and classifying one or more pixels as one of abnormal and normal, using the distance to the K samples and a label associated with the K samples, wherein a lesion likelihood index Lp for a pixel p of neighborhood Np is obtained from the K nearest samples (Sk)k=1 . . . K, with labels (lk)k=1 . . . K equal to 1 for lesions and −
1 for negative lesion confounders, by the following formula Lp=Σ
k=1Klkexp(−
∥
Np−
Sk∥
) and used to automatically classify the pixel p as abnormal or normal;and h. identifying one or more objects of interest in the reduced digital images from the classified pixels. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A method of identifying an object of interest in a plurality of digital images configured as an image group, wherein each of the digital images comprises a plurality of wavelength bands, the method comprising:
-
a) obtaining first samples of an intensity distribution of one or more objects of interest in the image group based upon one or more wavelength bands; b) obtaining second samples of the intensity distribution of confounder objects in the image group, at a frequency high enough to affect a pre-defined performance metric; c) transforming the first and second samples into an appropriate first space; d) performing dimensionality factor reduction on the transformed first and second samples, whereby the dimensionality reduction of the transformed first and second samples generates an object detector; e) transforming the image group into the first space; f) projecting the transformed image group into the reduced dimensionality space obtained in step d) to generate a reduced image group; g) classifying each pixel in the reduced image group with an appropriate neighborhood based on a comparison with the object detector, wherein the classification comprises; locating one or more K samples that minimize a distance to pixels in a pre-defined neighborhood; and classifying one or more pixels as one of abnormal and normal, using the distance to the K samples and a label associated with the K samples, wherein a lesion likelihood index Lp for a pixel p of neighborhood Np is obtained from the K nearest samples (Sk)k=1 . . . K, with labels (lk)k=1 . . . K equal to 1 for lesions and −
1 for negative lesion confounders, by the following formula Lp=Σ
k=1Klkexp(−
∥
Np−
Sk∥
) and used to automatically classify the pixel p as abnormal or normal; andh) automatically identifying one or more objects of interest in the reduced image group from abnormal pixels based upon the comparison with the object detector.
-
-
27. A system comprising:
-
a memory for storing digital images; and a processor in communication with the memory, the processor configured to; a) obtain first samples of an intensity distribution of one or more objects of interest in one or more of the digital images based upon one or more wavelength bands; b) obtain second samples of an intensity distribution of confounder objects in one or more of the digital images, at a predetermined frequency; c) transform the first and second samples into an appropriate first space; d) perform dimensionality factor reduction on the transformed first and second samples, whereby the dimensionality factor reduction of the transformed first and second samples generates an object detector; e) transform one or more of the digital images into the first space; f) perform dimensionality factor reduction on the transformed digital images, whereby the dimensionality reduction of the transformed digital images generates one or more reduced images; g) classify one or more pixels of the one or more reduced images based on a comparison with the object detector, wherein the classification comprises; locating one or more K samples that minimize a distance to pixels in a pre-defined neighborhood; and classifying one or more pixels as one of abnormal and normal, using the distance to the K samples and a label associated with the K samples wherein a lesion likelihood index Lp for a pixel p of neighborhood Np is obtained from the K nearest samples (Sk)k=1 . . . K, with labels (lk)k=1 . . . K equal to 1 for lesions and −
1 for negative lesion confounders, by the following formula Lp=Σ
k=1Klkexp(−
∥
Np−
Sk∥
) and used to automatically classify the pixel p as abnormal or normal; andh) identify one or more objects of interest in the reduced digital images from the classified pixels.
-
Specification