Digital makeup

• US 9,495,582 B2
• Filed: 12/04/2012
• Issued: 11/15/2016
• Est. Priority Date: 12/04/2011
• Status: Expired due to Fees

• Current Assignee: Digital Makeup Ltd.
• Original Assignee: Digital Makeup Ltd.
• Inventors: Guissin, Rami, Lavi, Eitan, Lifchitz, Guy
• Primary Examiner(s): Koziol, Stephen R
• Assistant Examiner(s): Lee, Jonathan S
• Application Number:
  US14/362,669
• Publication Number:
  US 20140328509A1
• Time in Patent Office: 1,442 Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/20024 Filtering details
• CPC Class Codes: G06T 2207/30201 Face
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships

Abstract

A system for processing an image of a human face, the system including a data processing and analyzing utility including a high-pass filtering module outputting a high-pass signal indicative of edges or contours of the face in the input image data; at least one band-pass filtering module outputting a bandpass signal indicative of low-contrast slowly-varying qualitative features of the face; a low-pass filtering module outputting a low-pass signal in which low-contrast regions are smoothed and high-contrast regions are preserved; a feature computation module calculating a localized feature of the image for a plurality of pixels of the image; a strength computation module determining a localized operand using the localized feature to determine a strength of filters to be used in the high-pass, band-pass, and low-pass filtering modules at the pixels, and transmitting the localized operand to the filtering modules; at least one transformation module altering a portion of the high-pass, the bandpass, or the low-pass signals; and an addition module adding the various signals together, thus yielding an output signal indicative of a characteristic of an altered image.

Claims

1. A system for processing an image of a human face, the system comprising a data processing and analyzing utility comprising:

a filtering system comprising a plurality of filtering modules, each filtering module being configured for receiving and processing data indicative of an input image data and generating an output signal, thereby providing a corresponding plurality of output signals indicative of different facial features, said plurality of filtering modules comprising;

a high pass filtering module comprising a first filtering utility configured for receiving the input image data, applying a first smoothing filter thereto, and outputting a first smoothed signal, and a first subtracting utility for subtracting the first smoothed signal from the input image data, thus yielding the high-pass signal indicative of at least one of edges and contours of the face in the input image data;

at least one bandpass filtering device comprising a second filtering utility, configured for receiving the input image data, applying a second smoothing filter thereto, and outputting a second smoothed signal, and a second subtracting utility for subtracting the second smoothed signal from the first smoothed signal, thus yielding at least one bandpass signal indicative of low-contrast slowly-varying qualitative features of the face; and

a low-pass filtering module configured for receiving the input image data and outputting a low-pass signal in which low-contrast regions are smoothed and high-contrast regions are preserved;

a feature computation module configured for receiving the input image data and calculating a localized feature of the image for a plurality of pixels of the image;

a strength computation module configured for receiving said localized feature from the feature computation module, using the localized feature for determining a localized operand to determine a strength for each filtering modules in said plurality of the filtering modules to be used at said plurality of the pixels, and transmitting the localized operand to each of said filtering modules;

at least one transformation module configured for receiving and altering at least a portion of at least one output signal of said plurality of output signals of the filtering modules; and

an addition module configured for receiving the output signals of the filtering modules after said at least portion of the at least one of said output signals has been altered, and for adding the received output signals together, thus yielding an output signal indicative of a characteristic of an altered image.

2. The system of claim 1, wherein the low-pass filtering module comprises the second filtering utility, the second smoothed signal being the low-pass signal.

5. The system of claim 1, wherein the strength computation module is configured for determining the localized operand to determine a strength of the smoothing filters at the plurality of the pixels, and transmitting the localized operand to the filtering utilities.

6. The system of claim 1, wherein at least one of the filtering utilities is configured for receiving an individual global parameter, and using the global parameter and the localized operand to calculate the strength of the respective smoothing filter.

7. The system of claim 1, wherein:

the strength computation module is configured for determining a plurality of localized operands, each localized operand corresponding to a respective filtering utility and being indicative of the strength of the respective smoothing filter; and

each filtering utility is configured for using the respective localized operand for determining the strength of the respective smoothing filter.

8. The system of claim 1, wherein at least one of the filtering utilities is configured for applying a multi-dimensional adaptive smoothing filter to a respective received signal.

9. The system of claim 1, wherein at least one of the filtering utilities is configured for applying a linear smoothing filter to a respective received signal.

10. The system of claim 9, wherein the first filtering utility is configured for applying the linear smoothing filter.

3. A system for processing an image of a human face, the system comprising a data processing and analyzing utility comprising:

a filtering system comprising a plurality of filtering modules, each filtering module being configured for receiving and processing data indicative of an input image data and generating an output signal, thereby providing a corresponding plurality of output signals indicative of different facial features, said plurality of filtering modules comprising;

a high pass filtering module comprising a first filtering utility configured for receiving the input image data, applying a first smoothing filter thereto, and outputting a first smoothed signal, and a first subtracting utility for subtracting the first smoothed signal from the input image data, thus yielding the high-pass signal indicative of at least one of edges and contours of the face in the input image data;

at least one bandpass filtering module comprising a second filtering utility configured for receiving the first smoothed signal, applying a second smoothing filter to the first smoothed signal, and outputting a second smoothed signal, and a second subtracting utility for subtracting the second smoothed signal from the first smoothed signal, thus yielding at least one bandpass signal indicative of low-contrast slowly-varying qualitative features of the face; and

a low-pass filtering module configured for receiving the input image data and outputting a low-pass signal in which low-contrast regions are smoothed and high-contrast regions are preserved;

a feature computation module configured for receiving the input image data and calculating a localized feature of the image for a plurality of pixels of the image;

a strength computation module configured for receiving said localized feature from the feature computation module, using the localized feature for determining a localized operand to determine a strength for each filtering modules in said plurality of the filtering modules to be used at said plurality of the pixels, and transmitting the localized operand to each of said filtering modules;

at least one transformation module configured for receiving and altering at least a portion of at least one output signal of said plurality of output signals of the filtering modules; and

an addition module configured for receiving the output signals of the filtering modules after said at least portion of the at least one of said output signals has been altered, and for adding the received output signals together, thus yielding an output signal indicative of a characteristic of an altered image.

4. The system of claim 3, wherein the low-pass filtering module comprises the second filtering unit, the second smoothed signal being the low-pass signal.

11. A system for processing an image of a human face, the system comprising a data processing and analyzing utility comprising:

a filtering system comprising a plurality of filtering modules, each filtering module being configured for receiving and processing data indicative of an input image data and generating an output signal, thereby providing a corresponding plurality of output signals indicative of different facial features, said plurality of filtering modules comprising;

a high pass filtering module configured for receiving the input image data indicative of a characteristic of the image and outputting a high-pass signal indicative of at least one of edges and contours of the face in the input image data;

an array of bandpass filtering modules, wherein each bandpass filtering module comprises;

a respective filtering utility configured for receiving said data indicative of the input image data and applying a respective smoothing filter thereto, and outputting a respective smoothed signal; and

a respective subtracting utility for subtracting the respective smoothed signal from a previously smoothed signal output by a preceding filtering utility comprised in a preceding filtering module, thus yielding a respective bandpass signal;

a low-pass filtering module configured for receiving the input image data and outputting a low-pass signal in which low-contrast regions are smoothed and high-contrast regions are preserved;

a feature computation module configured for receiving the input image data and calculating a localized feature of the image for a plurality of pixels of the image;

a strength computation module configured for receiving said localized feature from the feature computation module, using the localized feature for determining a localized operand to determine a strength for each filtering modules in said plurality of the filtering modules to be used at said plurality of the pixels, and transmitting the localized operand to each of said filtering modules;

at least one transformation module configured for receiving and altering at least a portion of at least one output signal of said plurality of output signals of the filtering modules; and

an addition module configured for receiving the output signals of the filtering modules after said at least portion of the at least one of said output signals has been altered, and for adding the received output signals together, thus yielding an output signal indicative of a characteristic of an altered image.

12. The system of claim 11, wherein the low-pass filtering module comprises a filtering utility being a part of a last bandpass filtering module in said array of the bandpass filtering modules, to generate a last smoothed signal being the low-pass signal.

13. A system for processing an image of a human face, the system comprising:

a first high-pass filtering module configured for receiving an input image data indicative of a characteristic of the image and outputting a first high-pass signal indicative of at least edges or contours of the face in the image;

a decimating module configured for receiving the input image data, and processing the received image data for reducing a size of the input image to produce a first decimated image, and outputting a decimated signal indicative of the first decimated image;

a filtering system, configured for receiving and processing the first decimated signal, and outputting a first output signal indicative of a first output image;

an interpolation module configured for receiving the first output signal, and interpolating the first output signal to output an interpolated signal indicative of an image having a size equal to the size of the input image; and

a first addition module configured for receiving the interpolated signal and the first high-pass signal, and adding the interpolated signal and the first high-pass signal together, thus yielding a second output signal indicative of a characteristic of an altered image.

14. The system of claim 13, comprising a first transformation module configured for altering at least a portion of the first high-pass signal to output a first transformed signal, the first addition module being configured for receiving the interpolated signal and the first transformed signal, and adding them together, thus yielding a second output signal indicative of a characteristic of an altered image.

15. The system of claim 13, wherein the first high-pass filtering module comprises:

a second interpolation module configured for receiving the decimated signal and interpolating the decimated signal to output a second interpolated signal, the first high-pass filtering module being configured for receiving the input image data and the second interpolated signal; and

a first subtracting unit configured for subtracting the second interpolated signal from the input image data thus yielding the first high-pass signal.

16. The system of claim 15, comprising a second decimating module configured for performing a second decimation on the first decimated signal to output a second decimated signal, which is received and processed by said filtering system to output the first output signal.

17. A method for processing an input signal indicative of a characteristic of an image of a human face, the method being implemented by a data processing and analyzing utility and comprising:

applying a first high-pass filter to input image data to generate a first high-pass signal indicative of at least edges or contours of the face in the image;

decimating the input image data to output a decimated signal indicative of an image of reduced size with respect to the input image;

processing the decimated signal to calculate a localized feature of the image for a plurality of pixels of the image;

using the localized feature for determining a localized operand;

applying a high-pass filter, at least one bandpass filter, and a low-pass filter to the decimated signal, thus generating a high-pass component signal, at least one bandpass component signal, and a low-pass component signal, such that a strength of each filter is dependent on the localized operand;

applying at least one transformation to at least one of the high-pass component signal, the bandpass component signal, and the low-pass component signal, for altering at least a portion of the at least one of the component signals;

adding the component signals after at least one of said component has been altered, thus yielding an output signal indicative of a characteristic of an altered image;

interpolating the output signal to output an interpolated signal indicative of an image having a size equal to the size of the input image; and

adding the first high-pass signal to the interpolated signal thus yielding a second output signal indicative of a characteristic of an altered image.

18. A non-transitory computer readable medium storing instructions that, when executed by a computer implementing a data processing and analyzing utility, cause the computer to carry out the following:

process input image data to calculate a localized feature of the image for a plurality of pixels of the image;

use the localized feature for determining a localized operand;

apply a high-pass filter, at least one bandpass filter, and a low-pass filter to the input image data, thus generating a high-pass component signal, at least one bandpass component signal, and a low-pass component signal, such that a strength of each filter is dependent on the localized operand, wherein

applying the high-pass filter comprises applying a first smoothing filter to the input image to provide a first smoothed signal, and subtracting the first smoothed signal from the input image data, thus yielding the high-pass component signal, and

applying the at least one bandpass filter comprises applying a second smoothing filter to the input image data to provide a second smoothed signal, and subtracting the second smoothed signal from the first smoothed signal, thus yielding the at least one bandpass component signal;

apply at least one transformation to at least one of the high-pass component signal, bandpass component signal, and low-pass component signal, for altering at least a portion of at least one of the component signals; and

add the component signals after at least one of said components has been altered, thus yielding an output signal indicative of a characteristic of an altered image.