# US 8,280,183 B2 · Noise robust contrast-enhancement engine employing IIR filters

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 8,280,183 B2
- Document ID: US8280183B2
- Status: Granted
- Kind: B2
- Application number: 12/133,854
- Filed: 2008-06-05
- Published: 2009-12-10
- Granted: 2012-10-02
- Earliest claimed: 2008-06-05
- Assignees: Texas Instruments Incorporated
- Inventors: Jeffrey Matthew Kempf; Christopher Paul Smearing
- Web page: https://patentgrep.com/patent/US8280183B2

## CPC

- G06T 5/94: based on local image properties, e.g. for local contrast enhancement
- G06T 5/20: using local operators
- G06T 5/40: using histogram techniques
- G06V 10/30: Noise filtering
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A noise robust contrast-enhancement engine utilizes low-pass infinite-impulse-response filters for enhancing the contrast while preventing noise amplification in video/image signals.

## Claims

### Claim 1

A method for enhancing a contrast of an image, comprising:
estimating a noise level of the image;
deriving a first signal from a pixel of the image based upon the noise floor using a first infinite-impulse-response filter, wherein the first signal represents a brightness of the image pixel;
deriving a second signal from the image pixel using a second infinite-impulse-response filter, wherein the second signal represents an illumination intensity of the image pixel; and
amplifying a difference between the first and second signals;
wherein the step of deriving a first signal further comprises applying the first infinite-impulse-filter to a luminance component of the image, wherein a set of coefficients of the first infinite-impulse-filter is determined by a temporal and spatial recursive process having a coefficient that is derived from a set of differences between the first signal and a luminance component of the image at a group of image pixel locations; and
wherein the step of deriving a second signal further comprises applying the second infinite-impulse-filter to a luminance component of the image, wherein a set of coefficients of the second infinite-impulse-filter is determined by a temporal and spatial recursive process having a coefficient that is derived from a variable and a chrominance component of the image, wherein said variable is derived from a recursive process for a group of image pixels.

### Claim 2 (depends on 1)

The method of claim 1, wherein the noise level is a noise floor of the image.

### Claim 3 (depends on 1)

The method of claim 1, wherein the step of amplifying a difference between the first and second signals further comprises amplifying the difference between the first and second signals with an amplification factor, wherein said amplification factor is determined based upon a probability that an image pixel is a skin-pixel.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
calculating an output luminance signal based upon the amplified difference and a luminance component in the input image; and
limiting the calculated luminance signal within a range of a color gamut.

### Claim 5

A method for enhancing a contrast in an image, comprising:
calculating a probability that an image pixel of the image is a skin pixel; and
enhancing a contrast of the image based upon the calculated probability, wherein the step of enhancing a contrast of the image further comprises:
estimating a noise floor of the image;
deriving a first signal from a pixel of the image based upon the noise floor using a first infinite-impulse-response filter, wherein the first signal represents a brightness of the image pixel; and wherein the step of deriving a first signal further comprises applying the first infinite-impulse-filter to a luminance component of the image, wherein a set of coefficients of the first infinite-impulse-filter is determined by a temporal and spatial recursive process having a coefficient that is derived from a set of differences between the first signal and a luminance component of the image at a group of image pixel locations;
deriving a second signal from the image pixel using a second infinite-impulse-response filter, wherein the second signal represents an illumination intensity of the image pixel; and
amplifying a difference between the first and second signals with an amplification factor that is determined based upon the calculated probability.

### Claim 6 (depends on 5)

The method of claim 5, further comprising:
applying a low-pass temporal filter to the determined noise floor, wherein said temporal filter comprises a set of filter-coefficients that are determined based upon a noise floor value from one image frame and another noise floor value from a different image frame.

### Claim 7

A method for enhancing a contrast in an image, comprising:
calculating a probability that an image pixel of the image is a skin pixel; and
enhancing a contrast of the image based upon the calculated probability, wherein the step of enhancing a contrast of the image further comprises:
estimating a noise floor of the image;
deriving a first signal from a pixel of the image based upon the noise floor using a first infinite-impulse-response filter, wherein the first signal represents a brightness of the image pixel;
deriving a second signal from the image pixel using a second infinite-impulse-response filter, wherein the second signal represents an illumination intensity of the image pixel; and wherein the step of deriving a second signal further comprises applying the second infinite-impulse-filter to a luminance component of the image, wherein a set of coefficients of the second infinite-impulse-filter is determined by a temporal and spatial recursive process having a coefficient that is derived from a variable and a chrominance component of the image, wherein said variable is derived from a recursive process for a group of image pixels; and
amplifying a difference between the first and second signals with an amplification factor that is determined based upon the calculated probability.

### Claim 8 (depends on 5)

The method of claim 5, further comprising:
calculating an output luminance signal based upon the amplified difference and a luminance component in the input image; and
limiting the calculated luminance signal within a range of a color gamut.

### Claim 9 (depends on 5)

The method of claim 5, wherein the step of deriving a second signal further comprises:
applying the second infinite-impulse-filter to a luminance component of the image, wherein a set of coefficients of the second infinite-impulse-filter is determined by a temporal and spatial recursive process having a coefficient that is derived from a variable and a chrominance component of the image, wherein said variable is derived from a recursive process for a group of image pixels.
