Face detection in digital images

Claim 3

The method as claimed in claim 1 , wherein said colour distribution model utilizes previously sampled data.

Claim 4

The method as claimed in claim 2 , wherein the colour distribution model is selected from a group consisting of models utilizing previously sampled data taken from images taken with use of a flash and images taken without use of a flash.

Claim 5

The method as claimed in claim 1 , wherein chromatic colours derived from RGB values are used in said testing step.

Claim 6

The method as claimed in claim 1 , wherein the facial feature analysis in said subjecting step is independent of facial colour.

Claim 7

The method as claimed in claim 5 wherein the facial feature analysis comprises edge detection.

Claim 8

The method as claimed in claim 6 , wherein the edge detection utilises an edge detection filter comprising a second derivative Gaussian function in a first direction substantially orthogonal to an edge to be detected and a Gaussian function in a second direction substantially orthogonal to the first direction.

Claim 9

The method as defined in claim 6 , wherein the facial feature analysis is carried out utilising the spatial relationship of detected edges.

Claim 10

The method as defined in claim 1 , wherein said segmenting step includes growing a region by determining the average colour of a group of pixels, determining the difference between the average colour of the group and the average colour of pixels adjacent to the group, and, if the colour difference is less than a predetermined threshold, adding the adjacent pixels to the group.

Claim 13

The apparatus as claimed in claim 10 , wherein said colour detecting means includes a storage means storing previously sampled data for use in said colour distribution model.

Claim 14

The apparatus as claimed in claim 11 , wherein the colour distribution model is selected from a group consisting of models utilizing previously sampled data taken from images taken with use of a flash and images taken without use of a flash.

Claim 15

The apparatus as claimed in claim 10 , further comprising chromatic colour calculations means to derive chromatic colour values from RGB values of the digital colour image.

Claim 16

The apparatus as claimed in claim 10 , wherein the facial feature analysis of said analysis means is independent of facial colour.

Claim 17

The apparatus as claimed in claim 14 , wherein said analysis means includes an edge detector to detect edges in the digital colour image.

Claim 18

The apparatus as claimed in claim 15 , wherein said analysis means has an edge detection filter comprising a second derivative Gaussian function in a first direction substantially orthogonal to an edge to be detected and a Gaussian function in a second direction substantially orthogonal to the first direction.

Claim 19

The apparatus as claimed in claim 15 or 16 , wherein said analysis means includes a spatial determinator to determine a spatial relationship of detected edges.

Claim 20

The apparatus as claimed in claim 10 , further comprising an accumulator means and a colour difference means to determine a difference between the average colour of a group of pixels and the average colour of pixels adjacent to that group, wherein if the colour difference is less than a predetermined threshold, the adjacent pixels are added to said accumulator means to thereby grow a region.

Claim 21

A computer readable medium incorporating a computer program product for detecting a face in a colour digital image captured by a camera, said computer program product including a sequence of computer implementable instructions for carrying out the steps of: segmenting the colour digital image into a plurality of regions each having a substantially homogenous colour; testing the colour of each region created in said segmenting step to determine those regions having predominantly skin colour, wherein said testing uses a colour distribution model that is dependent on settings of the camera at a time of capture of colour digital image; and subjecting only the regions determined in said testing step to facial feature analysis whereby the regions created in said segmenting step not having predominantly skin colour are not subjected to facial feature analysis.

Claim 22

A computer readable medium incorporating a computer program product for detecting a face in a colour digital image captured by a camera, said computer program product including a sequence of computer implementable instructions for carrying out the steps of: segmenting the colour digital image into a plurality of regions each having a substantially homogenous colour; testing the colour of each region created in said segmenting step to determine those regions having predominantly skin colour, wherein said testing uses a colour distribution model that is dependent on settings of the camera at a time of capture of colour digital image; and subjecting only the regions determined in said testing step to facial feature analysis whereby the regions created in said segmenting step not having predominantly skin colour are not subjected to facial feature analysis.

Claim 23

The computer readable medium as claimed in claim 19 , wherein said colour distribution model utilizes previously sampled data.

Claim 24

The computer readable medium as claimed in claim 20 , wherein the colour distribution model is selected from a group consisting of models utilizing previously sampled data taken from images taken with use of a flash and images taken without use of a flash.

Claim 25

The computer readable medium as claimed in claim 19 , wherein chromatic colours derived from RGB values are used in said testing step.

Claim 26

The computer readable medium as claimed in claim 19 , wherein the facial feature analysis in said subjecting step is independent of facial colour.

Claim 27

The computer readable medium as claimed in claim 23 , wherein the facial feature analysis comprises edge detection.

Claim 28

The computer readable medium as claimed in claim 24 , wherein the edge detection utilises an edge detection filter comprising a second derivative Gaussian function in a first direction substantially orthogonal to an edge to be detected and a Gaussian function in a second direction substantially orthogonal to the first direction.

Claim 29

The computer readable medium as claimed in claim 24 , wherein the facial feature analysis is carried out utilising a spatial relationship of detected edges.

Claim 30

The computer readable medium as claimed in claim 19 , wherein said segmenting step includes growing a region by determining the average colour of a group of pixels, determining the difference between the average colour of the group and the average colour of pixels adjacent to the group, and, if the colour difference is less than a predetermined threshold, adding the adjacent pixels to the group.

Claim 33

A computer readable medium incorporating a computer program product for detecting a face in a colour digital image captured by a camera, said computer program product including a sequence of computer implementable instructions for carrying out the steps of: segmenting the colour digital image into a plurality of regions each having a substantially homogenous colour; testing the colour of each region created in said segmenting step to determine those regions having predominantly skin color, wherein said testing uses a threshold altered by settings of the camera upon capture of the colour digital image; and subjecting only the regions determined in said testing step to facial feature analysis whereby the regions created in said segmenting step not having predominantly skin colour are not subjected to facial feature analysis.

Claim 34

A computer readable medium incorporating a computer program product for detecting a face in a colour digital image captured by a camera, said computer program product including a sequence of computer implementable instructions for carrying out the steps of: segmenting the colour digital image into a plurality of regions each having a substantially homogenous colour; testing the colour of each region created in said segmenting step to determine those regions having predominantly skin color, wherein said testing uses a threshold altered by settings of the camera upon capture of the colour digital image; and subjecting only the regions determined in said testing step to facial feature analysis whereby the regions created in said segmenting step not having predominantly skin colour are not subjected to facial feature analysis.