Multilevel chain-and-tree model for image-based decisions

• US 7,116,825 B2
• Filed: 03/22/2002
• Issued: 10/03/2006
• Est. Priority Date: 03/22/2002
• Status: Active Grant

• Current Assignee: Nikon Corporation
• Original Assignee: Chi-Chou Huang, Shih-Jong J. Lee
• Inventors: Huang, Chi-Chou, Lee, Shih-Jong J.
• Primary Examiner(s): DESIRE, GREGORY M
• Application Number:
  US10/104,669
• Publication Number:
  US 20030179936A1
• Time in Patent Office: 1,656 Days
• Field of Search: 382/226, 382/227, 382/160, 382/156, 382/209, 706/12, 706/20, 706/46
• US Class Current: 382/226
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06T 7/66 of image moments or centre of gravity

Abstract

A multilevel Chain-And-Tree model provides a framework for an image based decision system. The decision system enables separation of effects of defects within one component from other components within a common subject. The framework provides for linking of structure constraints of components of a common subject and for checking and resolving their consistency. The framework allows discrimination between subtle image changes and natural variations of the subject. The framework for standard data representation facilitates production process control.

Claims

1. An image-based decision method using a multilevel Chain-And-Tree model comprises the following steps:

(a) Input at least one reference multilevel CAT model basis of a subject selected from the set consisting of inspection specification, learning image, and application knowledge;

(b) Create a multilevel reference CAT model using at least one reference CAT model basis wherein the subject is represented by its components of geometric entities and their relations at multiple levels and the relations between components are represented as a chain or a tree link;

(c) Input a new image;

(d) Create a multilevel result CAT model having the same structure as the multilevel reference CAT model storing the measurement results of new image within the multilevel structure;

(e) Compare the multilevel reference CAT model and multilevel result CAT model to output a measurement result.

2. The method of claim 1 wherein the multilevel reference CAT model consists of the basic multilevel structure and attribute values of an ideal subject and its tolerance ranges wherein the attribute values include a level attribute, depth attribute, and component features.

3. The method of claim 1 wherein the multilevel CAT model contains at least one CAT chain wherein the relations between CAT chain nodes can be specified and wherein the relations are selected from the set consisting of Distance, Adjacency, InsideOutside, Parallelism, Perpendicularity, Concentricity, BetweenAngle, AreaRatio, and LengthRatio.

4. The method of claim 1 wherein the multilevel CAT model contains at least one CAT tree wherein the relations between CAT tree nodes can be specified and wherein the relations are selected from the set consisting of Distance, Adjacency, InsideOutside, Parallelism, Perpendicularity, Concentricity, BetweenAngle, AreaRatio, and LengthRatio.

5. The method of claim 1 wherein the multilevel CAT model contains at least one CAT node selected from the set consisting of chain root node, chain node, tree root node, and tree node.

6. An image-based decision method using a multilevel Chain-And-Tree model comprises the following steps:

(a) Input inspection specification;

(b) Input at least one learning image;

(c) Create a multilevel reference CAT model using the inputs wherein the subject is represented by its components of geometric entities and their relations at multiple levels and the relations between components are represented as a chain or a tree link;

(d) Create a multilevel processing CAT model using the inputs and the multilevel reference CAT model having the same structure as the multilevel reference CAT model storing the processing algorithm and sequence within the multilevel structure;

(e) Input a new image;

(f) Create a multilevel result CAT model using the processing CAT model and the new image having the same structure as the multilevel reference CAT model storing the measurement results of new image within the structure;

(g) Compare the multilevel reference CAT model and multilevel result CAT model to output at least one result.

7. The method of claim 6 wherein the multilevel reference CAT model consists of the basic multilevel structure and attribute values of an ideal subject wherein the attribute values include a level attribute, a depth attribute, and component features.

8. The method of claim 6 wherein the multilevel processing CAT model includes detection methods associated with the CAT component type for the detection of the CAT component.

9. The method of claim 6 wherein the CAT model contains at least one CAT chain wherein the relations between CAT chain nodes can be specified and wherein the relations are selected from the set consisting of Distance, Adjacency, InsideOutside, Parallelism, Perpendicularity, Concentricity, BetweenAngle, AreaRatio, and LengthRatio.

10. The method of claim 6 wherein the CAT model contains at least one CAT tree wherein the relations between CAT tree nodes can be specified and wherein the relations are selected from the set consisting of Distance, Adjacency, InsideOutside, Parallelism, Perpendicularity, Concentricity, BetweenAngle, AreaRatio, and LengthRatio.

11. The method of claim 6 wherein the CAT model contains at least one CAT node selected from the set consisting of chain root node, chain node, tree root node, and tree node.