Methods of improving detectors and classifiers using optimized stochastic resonance noise
First Claim
1. A method of improving the performance of a threshold-based detector or classifier, and increasing the probability of detecting or classifying at least one object in an image, said method comprising the steps of:
- a. obtaining image data by using an imaging system, wherein said image data comprises a first set of positive pixels and negative pixels;
b. calculating stochastic resonance noise with a processor by determining the stochastic resonance noise probability density function that maximizes the probability of detection while not increasing the probability of false alarm, and calculating the stochastic resonance noise data probability density function from a known probability density function for said image data, wherein the stochastic resonance noise probability density function can have the form of pnopt(n)=λ
δ
(n−
n1)+(1−
λ
)δ
(n−
n2), with values n1 and n2 equal to two delta function locations having probabilities of λ and
1−
λ
respectively, where n1 and n2 denote two discrete vectors, λ
denotes the probability of occurrence of said two vectors, and n is a randomization of said two discrete vectors added with the probabilities of λ and
1−
λ
, respectively; and
c. adding said stochastic resonance noise to said image data by using said processor, thereby improving the performance of a threshold-based detector or classifier and increasing the probability of detecting or classifying at least one object in an image.
1 Assignment
0 Petitions
Accused Products
Abstract
Apparatus and method for improving the performance of a threshold-based detector or classifier, or a generic detector or classifier and increasing the probability of detecting at least one object in an image using novel algorithms and stochastic resonance noise is provided, where a suitable dose of noise is introduced to the image data such that the performance of the above-referenced detectors or classifiers is improved without altering the detector'"'"'s or classifier'"'"'s parameters. Several stochastic resonance (SR) noise-based detection and classification enhancement schemes are presented. The SR noise-enhanced detection and classification schemes can improve any algorithms and systems. To implement these schemes, the only knowledge that is needed is the original input data (no matter 1D, 2D, 3D or others) and the output (detection results) of the existing algorithms and systems.
-
Citations
23 Claims
-
1. A method of improving the performance of a threshold-based detector or classifier, and increasing the probability of detecting or classifying at least one object in an image, said method comprising the steps of:
-
a. obtaining image data by using an imaging system, wherein said image data comprises a first set of positive pixels and negative pixels; b. calculating stochastic resonance noise with a processor by determining the stochastic resonance noise probability density function that maximizes the probability of detection while not increasing the probability of false alarm, and calculating the stochastic resonance noise data probability density function from a known probability density function for said image data, wherein the stochastic resonance noise probability density function can have the form of pnopt(n)=λ
δ
(n−
n1)+(1−
λ
)δ
(n−
n2), with values n1 and n2 equal to two delta function locations having probabilities of λ and
1−
λ
respectively, where n1 and n2 denote two discrete vectors, λ
denotes the probability of occurrence of said two vectors, and n is a randomization of said two discrete vectors added with the probabilities of λ and
1−
λ
, respectively; andc. adding said stochastic resonance noise to said image data by using said processor, thereby improving the performance of a threshold-based detector or classifier and increasing the probability of detecting or classifying at least one object in an image. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A method of improving the performance of a generic detector or classifier, and increasing the probability of detecting or classifying at least one object in an image, said method comprising the steps of:
-
a. obtaining image data by using an imaging system; b. calculating optimum stochastic resonance noise with a processor, wherein said optimum stochastic resonance noise comprises two-peak stochastic resonance noise having the following expression p n opt(n )=λ
δ
(n −
n 1)+(1−
λ
)δ
(n −
n 2), where λ and
1−
λ
are occurrence probabilities of the suitable N-dimensional vectorsn 1 andn 2, and 0≦
λ
≦
1, and where said optimum stochastic resonance noise includes three parameters,n 1,n 2 and λ
; andc. adding said optimum stochastic resonance noise to said image data by using said processor, thereby improving the performance of a detector and classifier and increasing the probability of detecting or classifying at least one object in an image. - View Dependent Claims (17, 18, 19, 20, 21, 22, 23)
-
Specification