Face recognition using pre-templates
First Claim
1. A method for facial recognition using a pre-template, comprising:
- detecting a face in a source image;
determining face information for the detected face;
cropping the source image around the detected face;
adjusting a compression level based on a contrast of at least a portion of the detected face using a processor;
encoding the cropped image based on the adjusted compression level;
storing the encoded cropped image and the face information in a storage medium;
obtaining an input image; and
in response to obtaining the input image;
obtaining the encoded cropped image from the storage medium,converting the encoded cropped image to a numerical representation of the detected face, andcomparing the numerical representation of the detected face with the input image.
2 Assignments
0 Petitions
Accused Products
Abstract
Disclosed herein is a method for creating a pre-template for use in facial recognition including detecting a face in a source image, determining face information for the detected face, cropping the source image around the detected face, adjusting a compression level based on a contrast of at least a portion of the detected face, encoding the cropped image based on the adjusted compression level, and storing the encoded cropped image and the face information in a storage medium, wherein the encoded cropped image is subsequently retrieved from the storage medium, decoded and converted to a numerical representation of the detected face to compare the numerical representation with a face in an input image.
14 Citations
40 Claims
-
1. A method for facial recognition using a pre-template, comprising:
-
detecting a face in a source image; determining face information for the detected face; cropping the source image around the detected face; adjusting a compression level based on a contrast of at least a portion of the detected face using a processor; encoding the cropped image based on the adjusted compression level; storing the encoded cropped image and the face information in a storage medium; obtaining an input image; and in response to obtaining the input image; obtaining the encoded cropped image from the storage medium, converting the encoded cropped image to a numerical representation of the detected face, and comparing the numerical representation of the detected face with the input image. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A method for creating a pre-template for use in facial recognition, comprising:
-
detecting a face in a source image; determining face information for the detected face; cropping the source image around the detected face; defining a facial boundary of the detected face in the cropped image; adjusting a compression level based on a contrast of at least a portion of the detected face using a processor, wherein the contrast is determined based on an area at least partially within the facial boundary, and adjusting the compression level includes adjusting a quantization matrix based on the contrast using at least one of the following equations;
Qij=α
+(i+j)β
γ
when i, j≠
0, or
Q00=α
+δ
γ
when i, j=0,wherein Q is the adjusted quantization matrix, i is a row in the adjusted quantization matrix, j is a column in the adjusted quantization matrix, γ
is a contrast-coefficient, representing the contrast measured within the facial boundary in the image, α
is a constant representing an expected level of white noise in the cropped image, β
is a constant providing frequency-dependent quantization, and δ
is a value for brightness-quantization;encoding the cropped image based on the adjusted compression level; and storing the encoded cropped image and the face information in a storage medium, wherein the encoded cropped image is subsequently retrieved from the storage medium, decoded and converted to a numerical representation of the detected face to compare the numerical representation with a face in an input image.
-
-
10. A method for creating a pre-template for use in facial recognition, comprising:
-
detecting a face in a source image; determining face information for the detected face; cropping the source image around the detected face; adjusting a compression level based on a contrast of at least a portion of the detected face using a processor by identifying a first area and at least a second area on the detected face, determining a contrast of the first area and a contrast of the second area, selecting a first quantization matrix to encode the first area of the detected face and a second quantization matrix to encode the second area of the detected face, the first quantization matrix having more quantization levels than the second quantization matrix, adjusting the first quantization matrix based on the contrast of the first area, and adjusting the second quantization matrix based on the contrast of the second area; encoding the cropped image based on the adjusted compression level; and storing the encoded cropped image and the face information in a storage medium, wherein the encoded cropped image is subsequently retrieved from the storage medium, decoded and converted to a numerical representation of the detected face to compare the numerical representation with a face in an input image.
-
-
11. An apparatus for creating a pre-template for use in facial recognition, comprising:
-
a memory; at least one processor configured to execute instructions stored in the memory to; detect a face in a source image; determine face information for the detected face; crop the source image around the detected face; identify a first area and at least a second area on the detected face; determine a contrast of the first area and a contrast of the second area; select a first quantization matrix to encode the first area of the detected face and a second quantization matrix to encode the second area of the detected face, the first quantization matrix having more quantization levels than the second quantization matrix adjust the first quantization matrix based on the contrast of the first area; adjust the second quantization matrix based on the contrast of the second area; encode the cropped image using the first quantization matrix and the second quantization matrix; and store the encoded cropped image and the face information in a storage medium, wherein the encoded cropped image is subsequently retrieved from the storage medium, decoded and converted to a numerical representation of the detected face to compare the numerical representation with a face in an input image. - View Dependent Claims (12, 13, 14)
-
-
15. A facial recognition system, comprising:
-
a storage medium; a first server configured to; detect a face in a source image; determine face information for the detected face; crop the source image around the detected face; identify a first area and at least a second area on the detected face; determine a contrast of the first area and a contrast of the second area; select a first quantization matrix to encode the first area of the detected face and a second quantization matrix to encode the second area of the detected face, the first quantization matrix having more quantization levels than the second quantization matrix adjust the first quantization matrix based on the contrast of the first area; adjust the second quantization matrix based on the contrast of the second area; encode the cropped image using the first quantization matrix and the second quantization matrix; store the encoded cropped image and the face information in the storage medium; and a second server configured to; retrieve the encoded cropped image from the storage medium; decode the encoded cropped image; convert the decoded cropped image to a numerical representation of the detected face using the face information; detect a face in an input image; and compare the numerical representation with the face in the input image.
-
-
16. A system, comprising:
-
one or more processors; and one or more memory devices for storing program instructions used by the one or more processors, wherein the program instructions, when executed by the one or more processors, cause the one or more processors to; detect a face in a source image; determine face information for the detected face; crop the source image around the detected face; adjust a compression level based on a contrast of at least a portion of the detected face using a processor; encode the cropped image based on the adjusted compression level; store the encoded cropped image and the face information in a storage medium; receive an input image; and in response to receiving the input image; retrieve the encoded cropped image from the storage medium, convert the encoded cropped image to a numerical representation of the detected face, and compare the numerical representation of the detected face with the input image. - View Dependent Claims (17, 18, 19, 20, 21)
-
-
22. A non-transitory computer-readable storage device including program instructions executable by one or more processors that, when executed, cause the one or more processors to perform operations, the operations comprising:
-
detecting a face in a source image; determining face information for the detected face; cropping the source image around the detected face; adjusting a compression level based on a contrast of at least a portion of the detected face using a processor; encoding the cropped image based on the adjusted compression level; storing the encoded cropped image and the face information in a storage medium; receiving an input image; and in response to receiving the input image; retrieving the encoded cropped image from the storage medium, converting the encoded cropped image to a numerical representation of the detected face, and comparing the numerical representation of the detected face with the input image. - View Dependent Claims (23, 24, 25, 26, 27)
-
-
28. A method for facial recognition using a pre-template, comprising:
-
selecting, by a processor, a compression level for an image based on a contrast of at least a portion of a detected face in the image; encoding the image based on the selected compression level; storing the encoded image in a storage medium; receiving an input image; and in response to receiving the input image; retrieving the encoded image from the storage medium; converting the encoded image to a numerical representation of the detected face, and comparing the numerical representation of the detected face with the input image. - View Dependent Claims (29, 30, 31, 32, 33)
-
-
34. A non-transitory computer-readable storage device including program instructions executable by one or more processors that, when executed, cause the one or more processors to perform operations, the operations comprising:
-
selecting a compression level for an image based on a contrast of at least a portion of a detected face in the image; encoding the image based on the selected compression level; storing the encoded image in a storage medium; receiving an input image; and in response to receiving the input image; retrieving the encoded image from the storage medium; converting the encoded image to a numerical representation of the detected face, and comparing the numerical representation of the detected face with the input image. - View Dependent Claims (35, 36, 37, 38, 39)
-
-
40. A system, comprising:
-
one or more processors; and one or more memory devices for storing program instructions used by the one or more processors, wherein the program instructions, when executed by the one or more processors, cause the one or more processors to; selecting a compression level for an image based on a contrast of at least a portion of a detected face in the image; encoding the image based on the selected compression level; storing the encoded image in a storage medium; receiving an input image; and in response to receiving the input image; retrieving the encoded image from the storage medium; converting the encoded image to a numerical representation of the detected face, and comparing the numerical representation of the detected face with the input image.
-
Specification