METHOD AND SYSTEM OF PERSON IDENTIFICATION BY FACIAL IMAGE
First Claim
1. A computer-based method of a person identification by a facial image comprising:
- localizing the facial image;
detecting coordinates of eye pupil centers on the facial image;
deriving a normalized facial image based on the facial image;
splitting the normalized facial image into rectangular blocks;
describing the rectangular blocks of the normalized facial image by a DCT (discrete cosine transform);
calculating distances between the normalized facial image and existing templates performed in two stages, the distances being weighted sums of shortest distances between each rectangular block of the normalized facial image and a set of rectangular blocks of the existing templates, wherein a small part of most different blocks are excluded from the weighted sum; and
recognizing the normalized facial image based on the distances calculated.
5 Assignments
0 Petitions
Accused Products
Abstract
A method to recognize a facial image is described. An input facial image is normalized by scaling and rotation angle using methods of eye pupil centers detection. The input facial image is further normalized by lighting intensity. Template images are obtained either by the processing of certain images taken from different face positions, or by a preliminary reconstruction of a 3D face model based on stereo-pair images. Using the 3D model template, facial images are generated at different rotation angles. Distances between the input facial image and the template image are calculated from the Discrete Cosine Transformation (DCT) features, defined by overlapped blocks of these images. The facial image is recognized based on these distances.
38 Citations
18 Claims
-
1. A computer-based method of a person identification by a facial image comprising:
-
localizing the facial image; detecting coordinates of eye pupil centers on the facial image; deriving a normalized facial image based on the facial image; splitting the normalized facial image into rectangular blocks; describing the rectangular blocks of the normalized facial image by a DCT (discrete cosine transform); calculating distances between the normalized facial image and existing templates performed in two stages, the distances being weighted sums of shortest distances between each rectangular block of the normalized facial image and a set of rectangular blocks of the existing templates, wherein a small part of most different blocks are excluded from the weighted sum; and recognizing the normalized facial image based on the distances calculated. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A machine-readable medium comprising instructions, which when implemented by one or more processors perform the following operations:
-
localize the facial image; detect coordinates of eye pupil centers on the facial image; derive a normalized facial image based on the facial image; split the normalized facial image into rectangular blocks; describe the rectangular blocks of the normalized facial image by a DCT (discrete cosine transform); calculate distances between the normalized facial image and existing templates performed in two stages, the distances being weighted sums of shortest distances between each rectangular block of the normalized facial image and a set of rectangular blocks of the existing templates, wherein a small part of most different blocks are excluded from the weighted sum; and recognize the normalized facial image based on the distances calculated.
-
-
18. A computer-based system of a person identification by a facial image comprising:
-
means for localizing the facial image; means for detecting coordinates of eye pupil centers on the facial image; means for deriving a normalized facial image based on the facial image; means for splitting the normalized facial image into rectangular blocks; means for describing the rectangular blocks of the normalized facial image by a DCT (discrete cosine transform); means for calculating distances between the normalized facial image and existing templates performed in two stages, the distances being weighted sums of shortest distances between each rectangular block of the normalized facial image and a set of rectangular blocks of the existing templates, wherein a small part of most different blocks are excluded from the weighted sum; and means for recognizing the normalized facial image based on the distances calculated.
-
Specification