Face recognition apparatus and method using physiognomic feature information

• US 9,990,538 B2
• Filed: 04/04/2016
• Issued: 06/05/2018
• Est. Priority Date: 01/05/2016
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Yoon, Ho-Sub, Ban, Kyu-Dae, Yoon, Young-Woo, Kim, Jae-Hong
• Primary Examiner(s): Patel, Jayesh A
• Application Number:
  US15/089,944
• Publication Number:
  US 20170193284A1
• Time in Patent Office: 792 Days
• Field of Search: None
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06V 40/165 using facial parts and geometric relationships
• CPC Class Codes: G06V 40/169 Holistic features and representations, i.e. based on the facial image taken as a whole
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A face recognition technology using physiognomic feature information, which can improve the accuracy of face recognition. For this, the face recognition method using physiognomic feature information includes defining standard physiognomic types for respective facial elements, capturing a facial image of a user, detecting information about facial elements from the facial image, and calculating similarity scores relative to the standard physiognomic types for respective facial elements of the user based on the facial element information.

Claims

1. A face recognition method using physiognomic feature information, comprising:

defining standard physiognomic types for respective facial elements;

capturing a facial image of a user;

performing basic face recognition using global features of a face in the facial image and detecting locations of the facial elements;

detecting information about the facial elements from the facial image;

normalizing a size of a face area in the facial image based on the facial elements;

calculating similarity scores relative to the standard physiognomic types for the respective facial elements of the user based on the facial elements;

assigning weights to the similarity scores and to results of the basic face recognition, respectively; and

determining similarities to facial data registered in advance in a face database (DB) based on the weighted similarity scores and the weighted results of the basic face recognition.

2. The face recognition method of claim 1, wherein the facial elements include at least one of eyes, a nose, a mouth, a chin, spots, and wrinkles, and

the standard physiognomic types are defined based on data including at least one of an interocular distance, presence/absence of double eyelids, a size of the eyes, a size and a shape of the nose, a shape of the chin, locations and shapes of the spots, and locations and shapes of the wrinkles.

3. The face recognition method of claim 1, wherein

the weights assigned to the similarity scores and to the results of the basic face recognition have values that are greater than 0 and less than 1, and

a total sum of the weights assigned to the similarity scores and to the results of the basic face recognition is 1.

4. The face recognition method of claim 2, wherein in response to detection of the facial elements correspond to the chin,

defining at least one of normal chin types including a rounded chin, a branchial chin, a pointed chin, an angular chin, and a square chin, and special chin types including a cleft chin, an asymmetric chin, and a double chin, as the standard physiognomic types for the chin based on preset criteria.

5. The face recognition method of claim 2, wherein in response to detection of the facial elements correspond to the spots or wrinkles,

defining the standard physiognomic types depending on locations, shapes, and a number of spots or wrinkles based on preset criteria.

6. The face recognition method of claim 1, further comprising:

defining standard physiognomic types of the face of the user based on the similarity scores relative to standard physiognomic types, which are calculated for respective facial elements of the user, and

determining similarities in response to detection of facial data stored in types identical to the defined standard physiognomic types in the face DB.

7. The face recognition method of claim 1, further comprising:

subsequent to detecting the information of the facial elements,

performing preprocessing for normalizing a size of a face in the facial image based on the information of the facial elements,

calculating similarity scores based on the information of the facial elements of the facial image on which the preprocessing having been performed.

8. The face recognition method of claim 1, further comprising:

capturing the facial image of the user using a normal camera and a 3D depth camera.