Active Shape Model for Vehicle Modeling and Re-Identification

• US 20080294401A1
• Filed: 05/19/2008
• Published: 11/27/2008
• Est. Priority Date: 05/21/2007
• Status: Active Grant

• Current Assignee: Siemens Healthineers AG (Siemens AG)
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Ramesh, Visvanathan, Genc, Yakup, Tsin, Yanghai
• Granted Patent: US 8,212,812 B2
• Time in Patent Office: Days
• US Class Current: 703/8
• CPC Class Codes: G06F 30/15 Vehicle, aircraft or watercraft design
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06T 19/20 Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
• CPC Class Codes: G06T 2210/44 Morphing
• CPC Class Codes: G06T 2219/2021 Shape modification
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/7553 based on shape, e.g. active shape models [ASM]
• CPC Class Codes: G06V 2201/08 Detecting or categorising vehicles

Abstract

A method for modeling a vehicle, includes: receiving an image that includes a vehicle; and constructing a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by: (a) taking a predetermined set of base shapes that are extracted from a subset of vehicles; (b) multiplying each of the base shapes by a parameter; (c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape; (d) fitting the vector to the vehicle in the image; and (e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized.

Claims

1. A method for modeling a vehicle, comprising:

receiving an image that includes a vehicle; and

constructing a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by;

(a) taking a predetermined set of base shapes that are extracted from a subset of vehicles;

(b) multiplying each of the base shapes by a parameter;

(c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape;

(d) fitting the vector to the vehicle in the image; and

(e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized.

2. The method of claim 1, wherein the base shapes are extracted by applying principle component analysis to vectors of landmark points, wherein each vector of landmark points represents a vehicle in the subset.

3. The method of claim 1, wherein a parameter is a linear interpolation coefficient.

4. The method of claim 1, wherein the parameters are found using a search algorithm.

5. The method of claim 4, wherein the search algorithm is a gradient descent.

6. A method for re-identifying a vehicle, comprising:

capturing an image that includes a first vehicle;

constructing a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by;

(a) taking a predetermined set of base shapes that are extracted from a subset of vehicles;

(b) multiplying each of the base shapes by a parameter;

(c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape;

(d) fitting the vector to the vehicle in the image; and

(e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized;

capturing an image that includes a second vehicle;

constructing a 3D model of the second vehicle, wherein the 3D model of the second vehicle is constructed by performing steps (a)-(e); and

determining if the 3D model of the first vehicle is the same as the 3D model of the second vehicle by comparing the 3D models to each other.

7. The method of claim 6, further comprising:

extracting color information from the vehicles in the images; and

applying the color information to the respective 3D models.

8. The method of claim 7, wherein the step of determining if the 3D model of the first vehicle is the same as the 3D model of the second vehicle further comprises:

comparing the colored 3D models to each other.

9. The method of claim 7, further comprising:

extracting environmental condition information of an area surrounding each of the vehicles in the images; and

applying the environmental condition information to the respective colored 3D models.

10. The method of claim 9, wherein when the environmental condition information is different, the method further comprises:

equalizing an effect of the environmental conditions on the colored 3D models.

11. The method of claim 6, further comprising:

extracting texture information from the vehicles in the images; and

applying the texture information to the respective 3D models.

12. The method of claim 11, wherein the step of determining if the 3D model of the first vehicle is the same as the 3D model of the second vehicle further comprises:

comparing the textured 3D models to each other.

13. The method of claim 11, further comprising:

extracting environmental condition information of an area surrounding each of the vehicles in the images; and

applying the environmental condition information to the respective textured 3D models.

14. The method of claim 13, wherein when the environmental condition information is different, the method further comprises:

equalizing an effect of the environmental conditions on the colored 3D models.

15. The method of claim 6, further comprising:

categorizing the 3D models.

16. A method for performing shape based control, comprising:

receiving an image that includes a vehicle; and

constructing a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by;

(a) taking a predetermined set of base shapes that are extracted from a subset of vehicles;

(b) multiplying each of the base shapes by a parameter;

(c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape;

(d) fitting the vector to the vehicle in the image; and

(e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized;

identifying a desired portion of the vehicle for further analysis; and

causing another image that includes the vehicle to be taken, wherein the image further includes an enhanced version of the desired portion.

17. The method of claim 16, wherein the images are captured from one or more than one camera.

18. A system for re-identifying a vehicle, comprising:

a first camera for capturing an image that includes a first vehicle;

a second camera for capturing an image that includes a second vehicle; and

a computer in communication with the cameras, the computer including a memory device for storing a program and a processor in communication with the memory device, the processor operative with the program to;

receive the image that includes the first vehicle;

construct a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by;

(a) taking a predetermined set of base shapes that are extracted from a subset of vehicles;

(b) multiplying each of the base shapes by a parameter;

(c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape;

(d) fitting the vector to the vehicle in the image; and

(e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized;

receive the image that includes the second vehicle;

construct a 3D model of the second vehicle, wherein the 3D model of the second vehicle is constructed by performing steps (a)-(e); and

determine if the 3D model of the first vehicle is the same as the 3D model of the second vehicle by comparing the parameters of the 3D models to each other.

19. The system of claim 18, wherein the first and second cameras are in different locations.

20. The system of claim 18, wherein the first and second cameras communicate with the computer over a wired or wireless network.

21. A computer readable storage medium storing instructions that, when executed by a computer, cause the computer to perform a method for modeling a vehicle, the method comprising:

receiving an image that includes a vehicle; and

constructing a three-dimensional (3D) model of the vehicle, wherein the 3D model is constructed by;

(a) taking a predetermined set of base shapes that are extracted from a subset of vehicles;

(b) multiplying each of the base shapes by a parameter;

(c) adding the resultant of each multiplication to form a vector that represents the vehicle'"'"'s shape;

(d) fitting the vector to the vehicle in the image; and

(e) repeating steps (a)-(d) by modifying the parameters until a difference between a fit vector and the vehicle in the image is minimized.