COMPUTER VISION CAD MODELS

• US 20100259537A1
• Filed: 10/10/2008
• Published: 10/14/2010
• Est. Priority Date: 10/12/2007
• Status: Abandoned Application

• Current Assignee: MvTec Software GmbH (MvTec LLC)
• Original Assignee: MvTec Software GmbH (MvTec LLC)
• Inventors: Ben-Himane, Selim, Hintestroisser, Stefan, Navab, Nassir
• Application Number:
  US12/682,199
• Publication Number:
  US 20100259537A1
• Time in Patent Office: Days
• US Class Current: 345/419
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06F 30/00 Computer-aided design [CAD]
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/809 of classification results, e.g. where the classifiers operate on the same input data

Abstract

The CV-CAD (computer vision-computer-aided design) model is an enhanced CAD (computer-aided design) model that integrates local and global computer vision data in order to represent an object not only geometrically but also in terms of computer vision. The CV-CAD model provides a scalable solution for intelligent and automatic object recognition, tracking and augmentation based on generic models of objects.

Claims

1. A method for creating a computer vision CAD model for use in object recognition or pose estimation from a standard CAD model of an object, the computer vision CAD model containing one or more agents that can be used for recognizing the object or parts of the object, said agents being attached to said object or to said parts of the object, comprising the steps of:

a) training said agents from the geometry of the object and the surface texture of the object,

b) storing said agents in the computer vision CAD model.

2. The method according to claim 1, wherein the agents comprise data that is used by object recognition or pose estimation algorithms that are defined outside the computer vision CAD model itself.

3. The method according to claim 1, wherein the agents comprise executable object recognition or pose estimation algorithms including data required for the object recognition or pose estimation algorithms.

4. The method of claim 1, wherein the surface texture of the object is stored in the standard CAD model.

5. The method of claim 1, wherein the surface texture of the object is obtained from one or more images of the object, additionally including the step of:

registering said one or more images of the object with the standard CAD model.

6. The method of claim 1, wherein the training the agents comprises selecting the best set of features and computer vision method to use for the object recognition or pose estimation.

7. The method of claim 1, wherein additionally geometric neighborhood relations and visibility information between the agents are trained.

8. The method of claim 1, wherein the agents are removable, addable, or replaceable by other agents if parts of the object are removed, added, or replaced.

9. The method of claim 1, wherein the agents are selected from any of the following agents:

point feature agent, edge-based agent, contour-based agent, color-based agent.

10. A method for using a computer vision CAD model for object recognition or pose estimation, the computer vision CAD model containing one or more agents that can be used for recognizing the object or parts of the object, said agents being attached to said object or to said parts of the object, comprising the steps of:

a) receiving the computer vision CAD model file,

b) generating executable instances of the agents that are stored in the computer vision CAD model file,

c) receiving an image,

d) recognizing the object described in the computer vision CAD model file or estimating its pose in said image by executing the agents, and

e) returning the object identity or its pose as well as the confidence of the object identity or pose.

11. The method of claim 10, wherein each agent returns hypotheses about the possible identity or pose of an object or object part.

12. The method of claim 11, wherein the a plurality of agents is used and the agents communicate with each other about possible object identities or pose hypotheses and confidence values of said possible object identities or said pose hypotheses.

13. The method of claim 12, wherein the agents select the best set of features and computer vision method to use for the object recognition or pose estimation with respect to said possible object identities, said pose hypotheses, geometric relationships, and an indication of visibility derived from the geometric relationships.

14. The method of claim 13, wherein the agents collaborate to derive a consistent hypothesis for the object identity or pose of the entire object stored in the computer vision CAD model.

15. A computer vision system comprising a computer vision CAD model for use in object recognition or pose estimation from a standard CAD model of an object, the computer vision CAD model containing one or more agents that can be used for recognizing the object or parts of the object, said agents being attached to said object or to said parts of the object, said system comprising:

means for training said agents from the geometry of the object and the surface texture of the object, and

means for storing said agents in the computer vision CAD model.

16. A computer vision system comprising a computer vision CAD model for object recognition or pose estimation, the computer vision CAD model containing one or more agents that can be used for recognizing the object or parts of the object, said agents being attached to said object or to said parts of the object, said system comprising:

means for receiving the computer vision CAD model file,

means for generating executable instances of the agents that are stored in the computer vision CAD model file,

means for receiving an image, means for recognizing the object described in the computer vision CAD model file or estimating its pose in said image by executing the agents, and

means for returning the object identity or its pose as well as the confidence of the object identity or pose.