Computer assisted domain specific entity mapping method and system

• US 20070078873A1
• Filed: 09/30/2005
• Published: 04/05/2007
• Est. Priority Date: 09/30/2005
• Status: Abandoned Application

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Sirohey, Saad, Avinash, Gopal, Weiner, Allison
• Application Number:
  US11/240,880
• Publication Number:
  US 20070078873A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/31 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/51 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 18/24 Classification techniques

Abstract

A technique for identifying, analyzing, structuring, mapping and classifying data entities is disclosed. A conceptual framework is established by a domain definition having an association list of attributes of interest. Data entities are accessed, analyzed, structured if appropriate, mapped and classified in accordance with the association list and attributes found in the entities, and in accordance with rules and algorithms for analyzing, recognizing and classifying the attributes. Various types of analysis may be performed following the classification. Searches and selection of the data entities may also be performed. Complex data entities may be processed, including text documents, image data, audio data, waveform data, and combinations of these.

Claims

1. A computer-implemented method for mapping data entities comprising:

defining a domain including a plurality of classification axes and a plurality of classification labels for each axis, and an association list of attributes associated with the axes and labels, at least one axis, label or attribute including an image feature, a waveform feature, an audio feature, a video feature or any combination thereof;

accessing a plurality of data entities potentially having attributes of interest;

identifying data entities having attributes corresponding to the axes and labels based upon the association list; and

classifying the identified data entities in accordance with the corresponding attributes.

2. The method of claim 1, wherein a label is defined by an image feature, a waveform feature, an audio feature or a video feature, and attributed of the label include variants of the label feature potentially included in the accessed data entities.

3. The method of claim 1, wherein multiple labels are defined for an axis by reference to an image feature, a waveform feature, an audio feature, or a video feature.

4. The method of claim 1, the classification includes a one-to-many mapping of a data entity to more than one label or axis.

6. The method of claim 1, wherein at least attributes associated with at least one label include attributes for at least two of textual features, image features, waveform features, audio features, and video features of data entities.

7. The method of claim 1, wherein the attributes encode features of medical images, and wherein the classification includes analysis of a disease state detectable from image data.

8. A computer-implemented method for mapping data entities comprising:

accessing a plurality of data entities potentially having attributes of interest; and

classifying the data entities based upon a domain definition including a plurality of classification axes, a plurality of classification labels for each axis, and an association list of attributes associated with the axes and labels to classify data entities having attributes corresponding to the axes and labels, wherein at least one axis, label or attribute including an image feature, a waveform feature, an audio feature, a video feature or any combination thereof.

9. The method of claim 8, wherein a label is defined by an image feature, a waveform feature, an audio feature or a video feature, and attributed of the label include variants of the label feature potentially included in the accessed data entities.

10. The method of claim 8, wherein multiple labels are defined for an axis by reference to an image feature, a waveform feature, an audio feature, or a video feature.

11. The method of claim 8, the classification includes a one-to-many mapping of a data entity to more than one label or axis.

12. The method of claim 8, wherein at least attributes associated with at least one label include attributes for at least two of textual features, image features, waveform features, audio features, and video features of data entities.

17. A computer-implemented method for mapping data entities comprising:

defining a domain including at least one classification axes and a plurality of classification labels the axis, and an association list of attributes associated with the labels, at least one label or attribute including an image feature, a waveform feature, an audio feature, a video feature or any combination thereof of interest for diagnosing a medical condition;

accessing a data entity potentially having attributes of interest;

identifying in the data entity attributes corresponding to the labels based upon the association list; and

classifying the data entity in accordance with the corresponding attributes for the diagnosis of the medical condition.

18. The method of claim 17, wherein the include attributes for at least two of textual features, image features, waveform features, audio features, and video features of data entities.

19. The method ofclaim 17, wherein defining the domain includes codifying an existing medical condition diagnosis standard.

20. The method of claim 17, wherein defining the domain includes defining the labels and attributes based upon a standard set of features having a known relationship to the medical condition.