Apparatus and methods for evaluating hyperdocuments using a trained artificial neural network

• US 7,403,929 B1
• Filed: 07/22/2005
• Issued: 07/22/2008
• Est. Priority Date: 07/23/2004
• Status: Active Grant

• Current Assignee: Ellis Robinson Giles, Vick Forrest Giles
• Original Assignee: Ellis Robinson Giles, Vick Forrest Giles
• Inventors: Giles, Vick Forrest, Giles, Ellis Robinson
• Primary Examiner(s): Holmes; Michael B.
• Application Number:
  US11/187,126
• Time in Patent Office: 1,096 Days
• Field of Search: 706/20, 707/101, 707/102
• US Class Current: 706/20
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations
• CPC Class Codes: G06N 3/02 Neural networks

Abstract

An embodiment of a computer implemented method for determining the disposition of a hyperdocument includes retrieving a hyperdocument from an information source, providing information about content of the hyperdocument to a trained artificial neural network (ANN), the ANN being capable of evaluating the information and providing results reflecting the evaluation, and determining the disposition of the hyperdocument based upon results of the ANN.

Claims

1. A computer implemented method for evaluating hyperdocument images, the method comprising:

retrieving a hyperdocument from an information source based upon a request from a requesting application;

providing information about one or more images of the hyperdocument to a trained artificial neural network, the trained artificial neural network being capable of evaluating the information and providing results reflecting the evaluation, wherein the information includes a pixel color value and a pixel RGB value for each image pixel;

converting each pixel color value into a double precision floating point value and mapping the pixel RGB value to a SVGA color by determining the closest color match; and

determining the disposition of the hyperdocument image based upon results of the trained artificial neural network, such disposition including at least one among modifying the hyperdocument and providing the modified hyperdocument to the requesting application or other destination, providing the original hyperdocument to the requesting application or other destination and providing a replacement hyperdocument to the requesting application or other destination.

2. The method of claim 1 wherein the artificial neural network is a feed forward artificial neural network having three layers with multiple input nodes, a single hidden layer having multiple hidden nodes and one output node, the input and output of the artificial neural network being represented by double precision floating point value numbers.

3. A method for filtering content of a hyperdocument the method comprising:

a user selecting criteria for allowable hyperdocuments;

retrieving a hyperdocument from an information source based upon a request from a requesting application;

providing information about content of the hyperdocument to a trained artificial neural network, the trained artificial neural network being capable of analyzing the information and providing results reflecting the analysis;

based upon the artificial neural network results, providing the hyperdocument to the requesting application if it matches the user selected criteria; and

comparing a hyperdocument'"'"'s source location to a list of disallowed information sources and if the hyperdocument'"'"'s source location is on the list, providing a designated alternate hyperdocument to the requesting application.

4. The method of claim 3 wherein the requesting application is web browser software and the hyperdocument is retrieved via the Internet.

5. The method of claim 3 further including a user designating at least one type of hyperdocument to be evaluated.

6. The method of claim 3 wherein the hyperdocument includes at least one image, further including converting hyperdocument image data into input values for the artificial neural network by at least one among scaling, color analysis and ordering of the image data.

7. The method of claim 6 wherein the artificial neural network evaluates the likelihood the hyperdocument image includes pornographic subject matter, the artificial neural network providing numeric output value that reflect the probability that the image includes pornographic subject matter.

8. The method of claim 7 further including

categorizing the numeric output value of the artificial neural network into one among multiple categories, the categories representing different levels of potential pornographic content ranging from YES to NO, and

comparing the category level of the numeric output value of the artificial neural network to a desired category level selected by the user.

9. The method of claim 3 further including allowing the user selected criteria to be on a scale of allowable hyperdocuments ranging from ALL to NONE.