Normalized image feature processing

• US 5,719,951 A
• Filed: 03/04/1993
• Issued: 02/17/1998
• Est. Priority Date: 07/17/1990
• Status: Expired due to Term

• Current Assignee: British Telecommunications PLC (BT Group PLC)
• Original Assignee: British Telecommunications PLC (BT Group PLC)
• Inventors: Shackleton, Mark Andrew, Welsh, William John
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US07/983,853
• Time in Patent Office: 1,811 Days
• Field of Search: 382/1, 382/2, 382/16, 382/30, 382/115, 382/117, 382/118, 382/209, 382/217, 382/199
• US Class Current: 382/118
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G07C 9/25 using biometric data, e.g. fingerprints, iris scans or voice recognition
• CPC Class Codes: H04N 19/00 Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
• CPC Class Codes: H04N 19/20 using video object coding

Abstract

A method of processing an image including the steps of: locating within the image the position of at least one predetermined feature; extracting from the image data representing each feature; and calculating for each feature a feature vector representing the position of the image data of the feature in an N-dimensional space, such space being defined by a plurality of reference vectors each of which is an eigenvector of a training set of like features in which the image data of each feature is modified to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of the feature, which step is carried out before calculating the corresponding feature vector.

Claims

1. A method of processing an image represented by an array of digital signals, said method comprising the steps off

locating within the image array of digital signals the position of at least one predetermined feature;

extracting image data from said image array of digital signals representing each said feature;

generating for each feature a feature vector digital signal representing the position of the image data of the feature in an N-dimensional space, N being an integer and said space being defined by a plurality of reference vector signals each of which is an eigenvector of a training set of images of like features; and

modifying the digital signal image data of each feature to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of said feature, which step is carried out before generating the feature vector signal for each feature.

5. A method according to claim 1 in which the training set of digital signal images of like features are modified to normalize the shape of each of the training set of digital signal images to reduce deviation from a predetermined standard shape of features and subsequently the reference vector eigenvectors of the training set of digital signal images are generated.

6. A method according to claim 1 wherein said locating step comprises:

locating a portion of the digital signal image by determining parameters of a closed curve arranged to lie adjacent a plurality of edge features of the digital signal image, said curve being constrained to exceed a minimum curvature and to have a minimum length compatible therewith.

7. A method as in claim 6 in which the curve is initially generated proximate the plurality of edge features of the digital signal image, and subsequently interactively re-generated to reduce its distance from said edge features.

8. A method according to claim 6 in which the located portion of the digital signal image is a face or a head of a person.

2. A method of processing an image represented by an array of digital signals, said method comprising the steps of:

locating within the image array of digital signals the position of at least one predetermined feature;

extracting image data from said image array of digital signals representing each said feature;

generating for each feature a feature vector digital signal representing the position of the image data of the feature in an N-dimensional space, N being an integer and said space being defined by a plurality of reference vector signals each of which is an eigenvector of a training set of images of like features; and

modifying the digital signal image data of each feature to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of said feature, which step is carried out before generating the feature vector signal for each feature,

wherein the step of modifying the digital signal image data of each feature uses a deformable template.

3. A method of processing an image represented by an array of digital signals, said method comprising the steps of:

locating within the image array of digital signals the position of at least one predetermined feature;

extracting image data from said image array of digital signals representing each said feature;

generating for each feature a feature vector digital signal representing the position of the image data of the feature in an N-dimensional space, N being an integer and said space being defined by a plurality of reference vector signals each of which is an eigenvector of a training set of images of like features; and

modifying the digital signal image data of each feature to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of said feature, which step is carried out before generating the feature vector signal for each feature,

wherein the step of locating within the digital signal image the position of at least one predetermined feature employs;

a first technique to provide a coarse estimation of position, and

a second, different, technique to improve upon the coarse estimation.

4. A method according to claim 3 wherein the second technique uses a deformable template.

9. A method of verifying the identity of the user of a data carrier, said method comprising the steps of:

generating a digital signal facial image of the user;

receiving the data carrier and reading therefrom identification data signals;

performing on the digital signal facial image the steps of;

(i) locating within the digital signal image the position of at least one predetermined feature;

(ii) extracting digital signal image data from said image representing each said feature;

(iii) generating for each feature a feature vector signal representing the position of the image data of the feature in an N-dimensional space, N being an integer and said space being defined by a plurality of reference vectors signals each of which is an eigenvector of a training set of images of like features; and

(iv) modifying the digital signal image data of each feature to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of said feature, which step is carried out before generating the feature vector for each feature;

comparing data signals representing each feature vector with the identification data signals; and

generating a verification signal in dependence upon the comparison.

10. Apparatus for verifying the identity of the user of a data carrier comprising:

means for generating a digital signal facial image of the user;

means for receiving the data carrier and reading therefrom identification data signals;

means for performing on the digital signal facial image the steps of;

(i) locating within the digital signal image and position of at least one predetermined feature;

(ii) extracting digital signal image data from said image representing each said feature;

(iii) generating for each feature a feature vector signal representing the position of the image data of the feature in an N-dimensional space, N being an integer and said space being defined by a plurality of reference vector signals each of which is an eigenvector of a training set of images of like features; and

(iv) modifying the digital signal image data of each feature to normalize the shape of each feature thereby to reduce its deviation from a predetermined standard shape of said feature, which step is carried out before calculating the feature vector for each feature; and

means for comparing data signals representing each feature vector with the identification data signals and generating a verification signal in dependence upon the comparison.

11. Apparatus as in claim 10 in which the means for generating a digital signal facial image of the user comprises a video camera, the output of which is connected to an AD convertor.