Automatic White Balance with Facial Color Features as Reference Color Surfaces

• US 20150326842A1
• Filed: 04/21/2015
• Published: 11/12/2015
• Est. Priority Date: 12/10/2014
• Status: Active Grant

• Current Assignee: Real Imaging Technology Co.Ltd
• Original Assignee: Real Imaging Technology Co.Ltd
• Inventors: Huai, Xiaoning
• Granted Patent: US 9,565,410 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30201 Face
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/90 Dynamic range modification of images or parts thereof
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 40/162 using pixel segmentation or colour matching
• CPC Class Codes: G08B 13/19621 Portable camera
• CPC Class Codes: G08B 13/19656 Network used to communicate with a camera, e.g. WAN, LAN, Internet
• CPC Class Codes: G08B 15/00 Identifying, scaring or incapacitating burglars, thieves or intruders, e.g. by explosives
• CPC Class Codes: G08B 25/006 Alarm destination chosen according to type of event, e.g. in case of fire phone the fire service, in case of medical emergency phone the ambulance
• CPC Class Codes: G08B 25/016 Personal emergency signalling and security systems emergency non-personal manually actuated alarm activators G08B25/12
• CPC Class Codes: H04N 23/10 for generating image signals from different wavelengths
• CPC Class Codes: H04N 23/60 Control of cameras or camera modules
• CPC Class Codes: H04N 23/61 based on recognised objects
• CPC Class Codes: H04N 23/611 where the recognised objects include parts of the human body
• CPC Class Codes: H04N 23/70 Circuitry for compensating brightness variation in the scene
• CPC Class Codes: H04N 23/71 Circuitry for evaluating the brightness variation
• CPC Class Codes: H04N 23/74 by influencing the scene brightness using illuminating means
• CPC Class Codes: H04N 23/88 for colour balance, e.g. white-balance circuits or colour temperature control
• CPC Class Codes: H04N 9/643 Hue control means, e.g. flesh tone control
• CPC Class Codes: H04N 9/646 : for image enhancement, e.g. vertical detail restoration, cross-colour elimination, contour correction, chrominance trapping filters
• CPC Class Codes: H04N 9/73 : Colour balance circuits, e.g. white balance circuits or colour temperature control camera processing pipelines for colour balance H04N23/88
• CPC Class Codes: H04W 4/90 : Services for handling of emergency or hazardous situations, e.g. earthquake and tsunami warning systems [ETWS]
• CPC Class Codes: H04W 76/50 : for emergency connections

Abstract

An automatic white balance method and an electronic media device incorporating the method deploy characterized individual human facial color features set of a user or users of the device as reference color surfaces in automatic white balance illuminant estimation and color adjustments, which could improve illuminant estimation accuracy, provide flexible color adjustment options and render the image color to each user'"'"'s own satisfaction taking into account user preferred rendition as target values through a human-machine interactive interface.

Claims

1. A method of automatic white balance of image and video media in an electronic media device, comprising:

acquiring one or a plurality of digital images of scenes under a sufficient color profile range of illumination conditions applicable to automatic white balance including a face of a user or users of the device, including capturing a digital image or images of a user or users using a lens and image sensor, or receiving a digital image or images of a user or users after capturing using a lens and image sensor, or a combination thereof;

acquiring the facial color features set of the user or users including the Intrinsic Values and Target Value of the color profiles of the skin tone portion, eye ball white and teeth by a color profile extraction process;

acquiring a digital image that needs automatic white balance and,

conducting face detection, segmentation and recognition in the digital image, and in case of a recognized face being identified belonging to the face of a user matching a user entry in the facial color features set;

conducting illuminant estimation based on segmented facial parts Features in the facial color features set resulting in C₂, and based on other parts of the image resulting in C₁,

and in final estimation C₃ taking into account both C₁ and C₂;

conducting automatic white balance gain adjustment and color enhancement based on the illuminant estimation results and the above acquired Facial Color Features Set of the user or users.

2. The method in claim 1, further comprising an integrated and a distributed implementation schemes, wherein the integrated implementation scheme accomplishes the method in one single device having at least a camera in front-side operational mode, including mobile phones, digital cameras or camcorders, laptop and desktop computers, other desk or mobile computing or communication devices and color imaging devices with built-in or connected camera or cameras, and the distributed implementation scheme accomplishes the method by more than one devices with at least one device having at least a camera in front-side operational mode, wherein capturing a digital image or images is carried out in a first device;

acquiring the facial color features set is carried out in a second device;

conducting illuminant estimation and automatic white balance gain adjustment and color enhancement is carried out in a third device, or a combination thereof;

3. The method in claim 1, wherein the acquiring the facial color features set of the automatic white balance method comprises:

capturing an image including a face of a user under a set of standard reference illuminations of known color profiles;

or capturing an image including a face of a user together with a standard reference grey chart in a variety of common shooting conditions and finding out the illumination profiles by analyzing the grey chart;

conducting face detection, segmentation and locating the contours of face skin, eyeball white and the teeth;

extracting the Intrinsic Values of the Facial Color Features of the face skin, the eyeball white and the teeth by averaging qualified pixels in each part after filtering using luminance threshold and histogram based color filter respectively;

extracting the Target Values of the Facial Color Features by applying the same procedure as above to the enhanced images of user choice or by user adjustment, or a combination thereof, through a user-device interface, together with the transform functions corresponding to the operations on captured image based on user input.

4. The method in claim 1, wherein the acquiring the facial color features set of the automatic white balance method comprises a machine learning mode, wherein the learning mode is activated or re-activated by a user or a system service routine, and one or more processors:

start illuminant estimation by C₁ in an unsupervised learning mode, and if a certain illuminant estimation C₁₀ has a confidence level w₁ higher than a threshold w_t, the one or more processors conduct face detection, segmentation recognition of the buffered image, and in case of a face of a user being identified matching a user entry in The facial color features set locate the contours of the face skin, the eyeball white and the teeth areas, extract the intrinsic values data of the facial color features of the face skin, the eyeball white and the teeth by averaging qualified pixels after filtering using a first luminance threshold filter and a second histogram based color filter and store the intrinsic values data in a candidate buffer;

if for the illuminant estimation C₁₀, the differences between the intrinsic values data stored in the candidate buffer are smaller than a threshold with a repetitive appearance probability P₁ bigger than a threshold P_1t, the illuminant estimation C₁₀ is considered accurate, and an average of the intrinsic values data within the difference threshold stored in the candidate buffer is computed and added into the related intrinsic values position of the user entry in the facial color features set;

in a supervised learning mode, the one or more processors proceed to display on a display device an image after automatic white balance adjustments adjusting the image color under current illumination estimated by C₁₀ to under a standard reference illumination or a popularly perceptually favored illumination condition followed by a color correction, and use a user-machine interactive interface to prompt for and acquire user'"'"'s opinion about the color rendition;

if the user confirms the color is satisfactory, extract the intrinsic values of the user under the estimated illumination C₁₀ using the same procedure described in the unsupervised learning mode and add the intrinsic values to the related intrinsic values position of the user entry in The facial color features set;

in a combinational mode, after the accuracy of an illuminant estimation C₁₀ passing through the statistical criteria in the unsupervised learning mode, conduct a further evaluation using the same procedure described in the supervised learning mode to confirm the accuracy of illuminant estimation C₁₀ before related intrinsic values data are added to The facial color features set;

acquire the target values data in the supervised learning mode by displaying enhanced images over or provide a mechanism for the user to adjust the captured image under the illumination estimated by C₁₀, or a combination thereof, and prompt for and acquire the input from the user;

extract the target values of the facial color features based on the input by the user using the same procedure described in the unsupervised learning mode, and add the target values together with the transform functions corresponding to operations on captured image based on user input to the related position of the user entry in the facial color features set.

complete the short term learning process after the acquired data in the user entry of The facial color features set cover sufficient range of color profiles as needed for automatic white balance, and illuminant estimation and automatic white balance adjustments could be carried out using the acquired data in The facial color features set.

5. The method in claim 1, wherein the illuminant estimation comprises:

using a (B/G, R/G) chromaticity coordinate system for illuminants with a black body radiation color profile,

calculating illuminant estimation for background C₁ first;

i_b denoting a pixel in the image background part, with its R, G and B components represented by r_ib, g_ib, b_ib respectively, the total number of pixels of image background portion is T₁, then an illuminant estimation for the background;

6. The method in claim 1, wherein the automatic white balance gain adjustment and color enhancement comprises:

applying the same gain offset for entire image, using C₃ as the preferred illuminant estimation and a pre-defined standard reference color profile such as D50 or D55 as the target illumination condition;

assuming C₃ being represented by a coordinate set (R1, G, B1) in RGB space, multiplying the R and B color component of each pixel of the image by a factor of G/R1 and G/B1 respectively;

applying color correction after automatic white balance gain adjustment by a 3*3 CCM matrix with constant coefficients, expressed as [R′

, G′

, B′

]=A*[R,G,B]^T.

7. The method in claim 1, wherein the automatic white balance gain adjustment and color enhancement comprises:

adjusting the facial parts and naked skin tone area from under current illuminant estimation C₂ to the user preferred target position defined as the Target Values in the user entry of the facial color features set, and color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position, by a first procedure comprising;

looking up for in the facial color features set and applying to the facial area and other detected naked skin area the same transform function as in the color profile extraction process transforming the original captured image to the enhanced image, with adjusted parameters in the function for the rendered image to match the target values corresponding to current illuminant estimation C₂;

adjusting the color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position;

smoothing the boundaries where heterogeneous adjustments occur;

or by a second procedure comprising;

Assuming B_current and G_current as the estimated skin tone under current illuminant estimation by C₂ and R_ref, B_ref and G_ref to be current Target Values for face skin corresponding to C₂, calculating

8. An electronic media device comprises:

one or more cameras with at least one camera working in a front-side operational mode;

one or more processors with one working as the system controller controlling system operation of the electronic media device;

non-transitory memory storage media containing instructions readable by the one or more processors;

memory storage media module accessible by the one or more processors to store data and execute the controlling instructions;

an apparatus with mechanisms to receive, store or output images, videos and other data;

a display device being able to display images, videos and graphics content;

a user controllable interactive interface to the electronic media device that takes in the user control input so that the system controller understands and properly responds to the user intention;

an automatic white balance control apparatus in the camera working in the front-side operational mode incorporating an automatic white balance method comprising;

acquiring one or a plurality of digital images of scenes under a sufficient color profile range of illumination conditions applicable to automatic white balance including a face of a user or users of the device, including capturing a digital image or images of a user or users using a lens and image sensor, or receiving a digital image or images of a user or users after capturing using a lens and image sensor, or a combination thereof;

acquiring the facial color features set of the user or users including the Intrinsic Values and Target Value of the color profiles of the skin tone portion, eye ball white and teeth by a color profile extraction process;

acquiring a digital image that needs automatic white balance and,

conducting face detection, segmentation and recognition in the digital image, and in case of a recognized face being identified belonging to the face of a user matching a user entry in the facial color features set;

conducting illuminant estimation based on segmented facial parts Features in the facial color features set resulting in C₂, and based on other parts of the image resulting in C₁, and in final estimation C₃ taking into account both C₁ and C₂;

conducting automatic white balance gain adjustment and color enhancement based on the illuminant estimation results and the above acquired Facial Color Features Set of the user or users.

9. The electronic media device in claim 8, wherein the automatic white balance method further comprising an integrated and a distributed implementation schemes, wherein the integrated implementation scheme accomplishes the method in one single device having at least a camera in front-side operational mode, including mobile phones, digital cameras or camcorders, laptop and desktop computers, other desk or mobile computing or communication devices and color imaging devices with built-in or connected camera or cameras, and the distributed implementation scheme accomplishes the method by more than one devices with at least one device having at least a camera in front-side operational mode, wherein capturing a digital image or images is carried out in a first device;

acquiring the facial color features set is carried out in a second device;

conducting illuminant estimation and automatic white balance gain adjustment and color enhancement is carried out in a third device, or a combination thereof;

10. The electronic media device in claim 8, wherein the acquiring the facial color features set of the automatic white balance method comprises:

capturing an image including a face of a user under a set of standard reference illuminations of known color profiles;

or capturing an image including a face of a user together with a standard reference grey chart in a variety of common shooting conditions and finding out the illumination profiles by analyzing the grey chart;

conducting face detection, segmentation and locating the contours of face skin, eyeball white and the teeth;

extracting the intrinsic values of the facial color features of the face skin, the eyeball white and the teeth by averaging qualified pixels in each part after filtering using luminance threshold and histogram based color filter respectively;

extracting the target values of the facial color features by applying the same procedure as above to the enhanced images of user choice or by user adjustment, or a combination thereof, through a user-device interface, together with the transform functions corresponding to the operations on captured image based on user input.

11. The electronic media device in claim 8, wherein the acquiring the facial color features set of the automatic white balance method further comprises a machine learning mode, wherein the learning mode is activated or re-activated by a user or a system service routine, and one or more processors:

start illuminant estimation by C₁ in an unsupervised learning mode, and if a certain illuminant estimation C₁₀ has a confidence level w₁ higher than a threshold w_t, the one or more processors conduct face detection, segmentation recognition of the buffered image, and in case of a face of a user being identified matching a user entry in The facial color features set locate the contours of the face skin, the eyeball white and the teeth areas, extract the intrinsic values data of the facial color features of the face skin, the eyeball white and the teeth by averaging qualified pixels after filtering using a first luminance threshold filter and a second histogram based color filter and store the intrinsic values data in a candidate buffer;

if for the illuminant estimation C₁₀, the differences between the intrinsic values data stored in the candidate buffer are smaller than a threshold with a repetitive appearance probability P₁ bigger than a threshold P_1t, the illuminant estimation C₁₀ is considered accurate, and an average of the intrinsic values data within the difference threshold stored in the candidate buffer is computed and added into the related intrinsic values position of the user entry in the facial color features set;

in a supervised learning mode, the one or more processors proceed to display on a display device an image after automatic white balance adjustments adjusting the image color under current illumination estimated by C₁₀ to under a standard reference illumination or a popularly perceptually favored illumination condition followed by a color correction, and use a user-machine interactive interface to prompt for and acquire user'"'"'s opinion about the color rendition;

if the user confirms the color is satisfactory, extract the intrinsic values of the user under the estimated illumination C₁₀ using the same procedure described in the unsupervised learning mode and add the intrinsic values to the related intrinsic values position of the user entry in The facial color features set;

in a combinational mode, after the accuracy of an illuminant estimation C₁₀ passing through the statistical criteria in the unsupervised learning mode, conduct a further evaluation using the same procedure described in the supervised learning mode to confirm the accuracy of illuminant estimation C₁₀ before related intrinsic values data are added to The facial color features set;

acquire the target values data in the supervised learning mode by displaying enhanced images over or provide a mechanism for the user to adjust the captured image under the illumination estimated by C₁₀, or a combination thereof, and prompt for and acquire the input from the user;

extract the target values of the facial color features based on the input by the user using the same procedure described in the unsupervised learning mode, and add the target values together with the transform functions corresponding to operations on captured image based on user input to the related position of the user entry in the facial color features set.

complete the short term learning process after the acquired data in the user entry of The facial color features set cover sufficient range of color profiles as needed for automatic white balance, and illuminant estimation and automatic white balance adjustments could be carried out using the acquired data in The facial color features set.

12. The electronic media device in claim 8, wherein the illuminant estimation of the automatic white balance method comprises:

using a (B/G, R/G) chromaticity coordinate system for illuminants with a black body radiation color profile,

calculating illuminant estimation for background C₁ first;

i_b denoting a pixel in the image background part, with its R, G and B components represented by r_ib, g_ib, b_ib respectively, the total number of pixels of image background portion is T₁, then an illuminant estimation for the background;

13. The electronic media device in claim 8, wherein the automatic white balance gain adjustment and color enhancement of the automatic white balance method comprises:

applying the same gain offset for entire image, using C₃ as the preferred illuminant estimation and a pre-defined color profile such as D50 or D55 as the standard target illumination condition;

assuming C₃ being represented by a coordinate set (R1, G, B1) in RGB space, multiplying the R and B color component of each pixel of the image by a factor of G/R1 and G/B1 respectively;

applying color correction after automatic white balance gain adjustment by a 3*3 CCM matrix with constant coefficients, expressed as [R′

, G′

, B′

]=A*[R, G, B]^T.

14. The electronic media device in claim 8, wherein the automatic white balance gain adjustment and color enhancement of the automatic white balance method comprises:

adjusting the facial parts and other detected naked skin area of a user from under current illuminant estimation C₂ to the user preferred target position defined as the Target Values in the user entry of the facial color features set, and color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position, by looking up for in the facial color features set and applying to the facial area and other detected naked skin area the same transform function as in the color profile extraction process transforming the original captured image to the enhanced images, with adjusted parameters in the function corresponding to current illuminant estimation C₂;

adjusting the color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position;

smoothing the boundaries where heterogeneous adjustments occur;

or by a second procedure comprising;

Assuming B_current and G_current as the estimated skin tone under current illuminant estimation by C₂ and R_ref, B_ref and G_ref to be current Target Values for face skin corresponding to C₂, calculating

15. The electronic media device in claim 8, wherein the automatic white balance method comprises a prioritized multiple user operational mode, wherein:

the OWNER or the administrator or a service routine of the device sets a priority rank number associated with each user entry in the facial color features set, so that the automatic white balance adjustment and color enhancement is biased to render image quality in the order favoring users with higher priority rank numbers when multiple faces of users are recognized in an image.

16. The electronic media device in claim 8, wherein the automatic white balance of the front-side camera comprising an operational mode to work concurrently with and support the automatic white balance illuminant estimation of the back-side camera, wherein:

the back-side camera requires the assistance of the illuminant estimation by the front-side camera through a service request to the system controller, and the system controller endorses the request when no resource or application conflict is found, and if the front-side camera is not in use by the user, launches the front-side camera, and the front-side camera conducts illuminant estimation and outputs the illuminant estimation results to the back-side camera;

the back-side camera uses the illuminant estimation results by the front-side camera as a reference for illuminant estimation by the back-side camera in situations comprising;

the fields of view of both cameras are both indoor or outdoor;

the distances of fields of view of both cameras are close to the device;

the color profile of light source detected by the front-side camera is the same or close to the illuminant estimation by the back-side camera;

the confidence level for illuminant estimation by the back-side camera is too low;

the illuminant estimation of both cameras are the same or close,

or a combination thereof.

18. The one or more non-transitory program storage devices in claim 16, wherein the method further comprising an integrated and a distributed implementation schemes, wherein the integrated implementation scheme accomplishes the method in one single device having at least a camera in front-side operational mode, including mobile phones, digital cameras or camcorders, laptop and desktop computers, other desk or mobile computing or communication devices and color imaging devices with built-in or connected camera or cameras, and the distributed implementation scheme accomplishes the method by more than one devices with at least one device having at least a camera in front-side operational mode, wherein capturing a digital image or images is carried out in a first device;

acquiring the facial color features set is carried out in a second device;

conducting illuminant estimation and automatic white balance gain adjustment and color enhancement is carried out in a third device, or a combination thereof.

19. The one or more non-transitory program storage devices in claim 16, wherein the acquiring the facial color features set of the automatic white balance method comprises:

capturing an image including a face of a user under a set of standard reference illuminations of known color profiles;

or capturing an image including a face of a user together with a standard reference grey chart in a variety of common shooting conditions and finding out the illumination profiles by analyzing the grey chart;

conducting face detection, segmentation and locating the contours of face skin, eyeball white and the teeth;

extracting the intrinsic values of the facial color features of the face skin, the eyeball white and the teeth by averaging qualified pixels in each part after filtering using luminance threshold and histogram based color filter respectively;

extracting the target values of the facial color features by applying the same procedure as above to the enhanced images of user choice or by user adjustment, or a combination thereof, through a user-device interface, together with the transform functions corresponding to the operations on captured image based on user input.

20. The electronic media device in claim 8, wherein the acquiring the facial color features set of the automatic white balance method further comprises a machine learning mode, wherein the learning mode is activated or re-activated by a user or a system service routine, and one or more processors:

start illuminant estimation by C₁ in an unsupervised learning mode, and if a certain illuminant estimation C₁₀ has a confidence level w₁ higher than a threshold w_t, the one or more processors conduct face detection, segmentation recognition of the buffered image, and in case of a face of a user being identified matching a user entry in The facial color features set locate the contours of the face skin, the eyeball white and the teeth areas, extract the intrinsic values data of the facial color features of the face skin, the eyeball white and the teeth by averaging qualified pixels after filtering using a first luminance threshold filter and a second histogram based color filter and store the intrinsic values data in a candidate buffer;

if for the illuminant estimation C₁₀, the differences between the intrinsic values data stored in the candidate buffer are smaller than a threshold with a repetitive appearance probability P₁ bigger than a threshold P_1t, the illuminant estimation C₁₀ is considered accurate, and an average of the intrinsic values data within the difference threshold stored in the candidate buffer is computed and added into the related intrinsic values position of the user entry in the facial color features set;

in a supervised learning mode, the one or more processors proceed to display on a display device an image after automatic white balance adjustments adjusting the image color under current illumination estimated by C₁₀ to under a standard reference illumination or a popularly perceptually favored illumination condition followed by a color correction, and use a user-machine interactive interface to prompt for and acquire user'"'"'s opinion about the color rendition;

if the user confirms the color is satisfactory, extract the intrinsic values of the user under the estimated illumination C₁₀ using the same procedure described in the unsupervised learning mode and add the intrinsic values to the related intrinsic values position of the user entry in The facial color features set;

in a combinational mode, after the accuracy of an illuminant estimation C₁₀ passing through the statistical criteria in the unsupervised learning mode, conduct a further evaluation using the same procedure described in the supervised learning mode to confirm the accuracy of illuminant estimation C₁₀ before related intrinsic values data are added to The facial color features set;

acquire the target values data in the supervised learning mode by displaying enhanced images over or provide a mechanism for the user to adjust the captured image under the illumination estimated by C₁₀, or a combination thereof, and prompt for and acquire the input from the user;

extract the target values of the facial color features based on the input by the user using the same procedure described in the unsupervised learning mode, and add the target values together with the transform functions corresponding to operations on captured image based on user input to the related position of the user entry in the facial color features set.

complete the short term learning process after the acquired data in the user entry of The facial color features set cover sufficient range of color profiles as needed for automatic white balance, and illuminant estimation and automatic white balance adjustments could be carried out using the acquired data in The facial color features set.

21. The one or more non-transitory program storage devices in claim 16, wherein the illuminant estimation of the method comprises:

using a (B/G, R/G) chromaticity coordinate system for illuminants with a black body radiation color profile,

calculating illuminant estimation for background C₁ first;

i_b denoting a pixel in the image background part, with its R, G and B components represented by r_ib, g_ib, b_ib respectively, the total number of pixels of image background portion is T₁, then an illuminant estimation for the background;

22. The one or more non-transitory program storage devices in claim 16, wherein the automatic white balance gain adjustment and color enhancement of the method comprises:

applying the same gain offset for entire image, using C₃ as the preferred illuminant estimation and a pre-defined color profile such as D50 or D55 as the standard target illumination condition;

assuming C₃ being represented by a coordinate set (R1, G, B1) in RGB space, multiplying the R and B color component of each pixel of the image by a factor of G/R1 and G/B1 respectively;

applying color correction after automatic white balance gain adjustment by a 3*3 CCM matrix with constant coefficients, expressed as [R′

, G′

, B′

]=A*[R, G, B]^T.

23. The one or more non-transitory program storage devices in claim 16, wherein the automatic white balance gain adjustment and color enhancement of the method comprises:

adjusting the facial parts and other detected naked skin area of a user from under current illuminant estimation C₂ to the user preferred target position defined as the Target Values in the user entry of the facial color features set, and color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position, by looking up for in the facial color features set and applying to the facial area and other detected naked skin area the same transform function as in the color profile extraction process transforming the original captured image to the enhanced images, with adjusted parameters in the function corresponding to current illuminant estimation C₂;

adjusting the color of the rest of the image from under current illuminant estimation C₁ to under a standard reference illumination position;

smoothing the boundaries where heterogeneous adjustments occur;

or by a second procedure comprising;

Assuming B_current and G_current as the estimated skin tone under current illuminant estimation by C₂ and R_ref, B_ref and G_ref to be current Target Values for face skin corresponding to C₂, calculating

17. One or more non-transitory program storage devices, containing instructions readable by one or more processors for causing the one or more processors to perform a method of automatic white balance, comprising:

acquiring one or a plurality of digital images of scenes under a sufficient color profile range of illumination conditions applicable to automatic white balance including a face of a user or users of the device, including capturing a digital image or images of a user or users using a lens and image sensor, or receiving a digital image or images of a user or users after capturing using a lens and image sensor, or a combination thereof;

acquiring the facial color features set of the user or users including the Intrinsic Values and Target Value of the color profiles of the skin tone portion, eye ball white and teeth by a color profile extraction process;

acquiring a digital image that needs automatic white balance and,

conducting face detection, segmentation and recognition in the digital image, and in case of a recognized face being identified belonging to the face of a user matching a user entry in the facial color features set;

conducting illuminant estimation based on segmented facial parts Features in the facial color features set resulting in C₂, and based on other parts of the image resulting in C₁, and in final estimation C₃ taking into account both C₁ and C₂;

conducting automatic white balance gain adjustment and color enhancement based on the illuminant estimation results and the above acquired Facial Color Features Set of the user or users.