US 2012/0070041 A1Published application

System And Method For Face Verification Using Video Sequence

AssigneeSeiko Epson Corporation

InventorJie Wang

Life of this patent

Estimating the term…

Loading patent…

Abstract

Face verification is performed using video data. The two main modules are face image capturing and face verification. In face image capturing, good frontal face images are captured from input video data. A frontal face quality score discriminates between frontal and profile faces. In face verification, a local binary pattern histogram is selected as the facial feature descriptor for its high discriminative power and computational efficiency. Chi-Square (χ2) distance between LBP histograms from two face images are then calculated as a face dissimilarity measure. The decision whether or not two images belong to the same person is then made by comparing the corresponding distance with a pre-defined threshold. Given the fact that more than one face images can be captured per person from video data, several feature based and decision based aggregators are applied to combine pair-wise distances to further improve the verification performance.

Claims20

Dependent claims

A method for determining if an image of a face is a face candidate, where a face rectangle is defined as {xf,yf,hf,wf}, where {xf,yf} are the coordinates of a left top point, {hf,wf} are the height and width of the face rectangle, respectively, left and right eyes of the face are defined as {xle,yle,hle,wle} and {xre,yre,hre,wre} respectively, with {xle,yle}, {xre,yre} being left top points of the coordinates of the left and right eyes, respectively, and {hle,wle}, {hre,wre} being the height and width of the left and right eyes, respectively, and {hI,wI} denotes the height and width of the image, respectively, comprising:

capturing an image of a face with an image capture device;
and
using a processor to determine if the face is a face candidate by examining the face for the following criteria:
(1) the face should appear in the middle of the image, such that xf≧θl×wI, xf+wf≦θr×wI, yf≦θt×hI, yf+hf≦θb×hI, where 0≦θl, θr, θt, θb≦1 and θl, θr, θt, θb are four pre-defined thresholds;
(2) the face should be sufficiently large, such that min(hf,wf)≧θs×min(hI,wI), θs is a pre-defined threshold;
(3) the left and right eyes should be located in a horizontal line, such that |yle−yre|≦max(min(hle,hre)×γ1,3) where γ1 is a pre-defined threshold;
(4) the left and right eyes should have similar size, such that |wle−wre|≦max(min(wle,wre)×γ2,3), where γ2 is a pre-defined threshold;
(5) a distance between the left and eyes should be within a range, such that wf×γ3<|xlcc−xrcc|<wf×γ4, where xlcc, xrcc are x-coordinates of the left and right eye rectangle center points, respectively, and γ3 and γ4 are pre-defined thresholds;
(6) a size of the left and right eyes should be within a range, i.e., min(wle,wre)≧wf×γ5, max(wle,wre)≦wf×γ6, where γ5 and γ6 are pre-defined thresholds;
and (7) the left and right eyes should be located in the middle of the face, i.e., xf+wf−xre−wre≧wre×γ7, xle−xf≧wle×γ8, where γ7 and γ8 are pre-defined thresholds;
and outputs to an output device a result that the image of a face is a face candidate when the criteria (1) through (7) are met.

8 more elements

A device for determining if an image of a face is a face candidate, where a face rectangle is defined as {xf,yf,hf,wf}, where {xf,yf} are the coordinates of a left top point, {hf,wf} are the height and width of the face rectangle, respectively, left and right eyes of the face are defined as {xle,yle,hle,wle} and {xre,yre,hre,wre} respectively, with {xle,yle}, {xre,yre} being the coordinates of the left and right eyes, respectively, and {hle,wle}, {hre,wre} being the height and width of the left and right eyes, respectively, and {hI,wI} denotes the height and width of the image, respectively, comprising:

an image capture device that captures an image of a face;
and
a processor that determines if the face is a face candidate by examining the face for the following criteria:
(1) the face should appear in the middle of the image, such that xf≧θl×wI, xf+wf≦θr×wI, yf≦θt×hI, yf+hf≦θb×hI, where 0≦θl, θr, θt, θb≦1 and θl, θr, θt, θb are four pre-defined thresholds;
(2) the face should be sufficiently large, such that min(hf,wf)≧θs×min(hI,wI), θs is a pre-defined threshold;
(3) the left and right eyes should be located in a horizontal line, such that |yle−yre|≦max(min(hle,hre)×γ1,3) where γ1 is a pre-defined threshold;
(4) the left and right eyes should have similar size, such that |wle−wre|≦max(min(wle,wre)×γ2,3), where γ2 is a pre-defined threshold;
(5) a distance between the left and eyes should be within a range, such that wf×γ3<|xlcc−xrcc|<wf×γ4, where xlcc, xrcc are x-coordinates of the left and right eye rectangle center points, respectively, and γ3 and γ4 are pre-defined thresholds;
(6) a size of the left and right eyes should be within a range, i.e., min(wle,wre)≧wf×γ5, max(wle,wre)≦wf×γ6, where γ5 and γ6 are pre-defined thresholds;
and (7) the left and right eyes should be located in the middle of the face, i.e., xf+wf−xre−wre≧wre×γ7,xle−xf≧wle×γ8, where γ7 and γ8 are pre-defined thresholds;
and an output device that receives a result from the processor that the image of a face is a face candidate when the criteria (1) through (7) are met.

8 more elements

One or more tangible computer-readable media having computer-readable instructions thereon, which, when executed by a processor determine if an image of a face is a face candidate, where a face rectangle is defined as {xf,yf,hf,wf}, where {xf,yf} are the coordinates of a left top point, {hf,wf} are the height and width of the face rectangle, respectively, left and right eyes of the face are defined as {xle,yle,hle,wle} and {xre,yre,hre,wre} respectively, with {xle,yle}, {xre,yre} being the coordinates of the left and right eyes, respectively, and {hle,wle}, {hre,wre} being the height and width of the left and right eyes, respectively, and {hI,wI} denotes the height and width of the image, respectively, wherein:

the processor determines if the face is a face candidate by examining the face for the following criteria:
(1) the face should appear in the middle of the image, such that xf≧θl×wI, xf+wf≦θr×wI, yf≦θt×hI, yf+hf≦θb×hI, where 0≦θl, θr, θt, θb≦1 and θl, θr, θt, θb are four pre-defined thresholds;
(2) the face should be sufficiently large, such that min(hf,wf)≧θs×min(hI,wI), θs is a pre-defined threshold;
(3) the left and right eyes should be located in a horizontal line, such that |yle−yre|≦max(min(hle,hre)×γ1,3) where γ1 is a pre-defined threshold;
(4) the left and right eyes should have similar size, such that |wle−wre|≦max(min(wle,wre)×γ2,3), where γ2 is a pre-defined threshold;
(5) a distance between the left and eyes should be within a range, such that wf×γ3<|xlcc−xrcc|<wf×γ4, where xlcc, xrcc are x-coordinates of the left and right eye rectangle center points, respectively, and γ3 and γ4 are pre-defined thresholds;
(6) a size of the left and right eyes should be within a range, i.e., min(wle,wre)≧wf×γ5, max(wle,wre)≦wf×γ6, where γ5 and γ6 are pre-defined thresholds;
and (7) the left and right eyes should be located in the middle of the face, i.e., xf+wf−xre−wre≧wre×γ7, xle−xf≧wle×γ8, where γ7 and γ8 are pre-defined thresholds;
and the processor outputs to an output device a result that the image of a face is a face candidate when the criteria (1) through (7) are met.

6 more elements

Dependencies are inferred from the claim text.

Facts

Application: 12/883,931Filed 16 Sep 2010
Published: 22 Mar 2012
Term ends: Not estimatedThe source records don’t support an estimate.

Classification6 codes

G06F 21/32
User authentication: using biometric data, e.g. fingerprints, iris scans or voiceprints
G06V 40/165
Detection; Localisation; Normalisation: using facial parts and geometric relationships
G06V 40/171
Feature extraction; Face representation: local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
G06V 10/467
Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient…:…