SYSTEM AND METHOD FOR USING THREE DIMENSIONAL INFRARED IMAGING TO IDENTIFY INDIVIDUALS

• US 20100189313A1
• Filed: 04/15/2008
• Published: 07/29/2010
• Est. Priority Date: 04/17/2007
• Status: Active Grant

• Current Assignee: Francine J. Prokoski
• Original Assignee: Francine J. Prokoski
• Inventors: Prokoski, Francine J.
• Granted Patent: US 8,494,227 B2
• Time in Patent Office: Days
• US Class Current: 382/118
• CPC Class Codes: A61B 5/0064 Body surface scanning
• CPC Class Codes: A61B 5/0086 using infrared radiation
• CPC Class Codes: A61B 5/015 By temperature mapping of body part
• CPC Class Codes: A61B 5/1077 Measuring of profiles
• CPC Class Codes: A61B 5/411 Detecting or monitoring allergy or intolerance reactions to an allergenic agent or substance
• CPC Class Codes: A61B 5/415 the glands, e.g. tonsils, adenoids or thymus
• CPC Class Codes: A61B 5/418 lymph vessels, ducts or nodes
• CPC Class Codes: A61B 5/445 Evaluating skin irritation or skin trauma, e.g. rash, eczema, wound, bed sore
• CPC Class Codes: A61B 5/4519 Muscles A61B5/389, A61B5/224 take precedence
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G01K 13/20 Clinical contact thermometers for use with humans or animals
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G16H 30/20 for handling medical images, e.g. DICOM, HL7 or PACS
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

Calibrated infrared and range imaging sensors are used to produce a true-metric three-dimensional (3D) surface model of any body region within the fields of view of both sensors. Curvilinear surface features in both modalities are caused by internal and external anatomical elements. They are extracted to form 3D Feature Maps that are projected onto the skin surface. Skeletonized Feature Maps define subpixel intersections that serve as anatomical landmarks to aggregate multiple images for models of larger regions of the body, and to transform images into precise standard poses. Features are classified by origin, location, and characteristics to produce annotations that are recorded with the images and feature maps in reference image libraries. The system provides an enabling technology for searchable medical image libraries.

Claims

1. A method for identifying a person, comprising:

generating a body map of one or more body segments of the person;

comparing the body map of said one or more body segments to body maps of corresponding segments of known persons; and

applying a threshold test to determine whether one or more body maps of corresponding segments of known persons is a match.

2. A method of identification as in claim 1, wherein each body map is generated by:

collecting simultaneous images of the body segment using a plurality of imaging devices;

combining corrected and overlaid images of the infrared imager and the range imager to produce a 3D infrared model;

processing the range image to extract a curvilinear feature map of external anatomy;

processing the infrared image to extract a curvilinear feature map of internal anatomy;

skeletonizing the respective curvilinear feature maps;

producing skeleton node maps containing intersection and branch locations of said curvilinear features;

labeling each node according to a standard directory description of intersecting or branching anatomical features;

forming a layered composite image of the infrared, range, and visual images, plus their feature maps, plus their skeletonized feature maps, plus their node maps;

selecting nodes corresponding to three reference points designated for said body segment;

rotating the composite image in three-dimensional space such that the three reference points define a two dimensional (2D) image plane, said 2D image plane being a standard pose for said body segment;

storing said rotated standardized composite image as a body map.

3. A method of identification as in claim 1, wherein the body map is of the face area, further comprising:

performing locally adaptive filtering on the infrared layer of the body map composite to enhance visualization of internal anatomical structures including blood vessels;

thresholding the enhanced infrared image to produce binary curvilinear features;

skeletonizing the features to produce single-pixel line widths; and

calculating measurements of skeletonized features including number of features, total length of line features, distribution of line feature lengths, and distribution of feature angular orientation,

wherein said set of feature measurements from the facial body map of the person and the known person are used in the threshold test.

4. A method of identification as in claim 3, further comprising:

creating a node map showing intersections and branch points in the facial body map;

determining distribution of nodes for the facial body map in a standard pose (frontal or profile) with respect to bilateral symmetry, number and type of node in each segment of the face;

determining (x,y,z) location relative to a body-centric coordinate system and type for each node;

wherein said set of node measurements is used in the threshold test.

5. A method of identification as in claim 4, wherein said set of node measurements includes vector angles of features intersecting or branching to form each node location.

6. A method of identification as in claim 4, wherein said set of node measurements includes range value at each node location.

7. A method for identification of a person by comparison against a database of known persons, comprising:

generating a body map of the face area of the person;

standardizing pose for the generated facial body map and corresponding facial body maps in the database of known persons;

producing a two dimensional (2D) FaceMap from the generated facial body map;

comparing the 2D FaceMap against the database of similarly obtained FaceMaps of known persons;

computing a correlation between FaceMaps by computing differences in characterization for all infrared features that have at least a specified number of pixels in common locations for the two FaceMaps;

rank order all comparisons;

apply a threshold to the best match to determine confidence that it is a correct match.

8. A method of identification as in claim 7, further comprising:

generating a node map from the FaceMap;

characterizing features and nodes on the nodemap;

generating a FaceCode consisting of a list of coordinates of each node plus (x,y,z) coordinates of all pixels contained in infrared features.

9. The method of identification of claim 8, wherein the FaceCode contains a subset of the list of coordinates of each node plus (x,y,z) coordinates of all pixels contained in infrared features.

10. A method for identification of a person by generation of a FaceCode representing differences between his FaceMap and a reference FaceMap, comprising:

generating a body map of the face area in standard pose;

characterizing features and nodes on the generated FaceMap;

comparing the characterizations to those associated with a reference FaceMap;

generating a FaceCode consisting of differences between selected characteristics of the person'"'"'s FaceMap and the reference FaceMap.

11. A method for comparing infrared facial images to a database of visual facial images, comprising:

generating a FaceMap;

for each visual facial image in the database, rotating the FaceMap to match pose of the visual image;

producing a 2D FaceMap in the matching pose;

comparing an external feature layer of the 2D FaceMap to feature edges in the visual image;

computing the percentage of coincident pixels;

discarding images below threshold;

overlaying IR features on the visual image and detecting whether internal IR features intersect edge feature in the visual image and, if so, discarding that visual image as a potential match; and

rank ordering remaining potential matches according to percentage of coincident pixels.