Neural-network-based method of image compression

• US 6,798,914 B1
• Filed: 05/05/2000
• Issued: 09/28/2004
• Est. Priority Date: 05/07/1999
• Status: Expired due to Fees

• Current Assignee: Galaxy IP Source, LLC
• Original Assignee: Galaxy IP Source, LLC
• Inventors: Abraham, Gerald, Nanni, Richard A.
• Primary Examiner(s): Couso, Jose L.
• Application Number:
  US09/565,372
• Time in Patent Office: 1,607 Days
• Field of Search: 382/155-162, 382/164-166, 382/181, 382/187, 382/226, 382/227, 382/232, 382/233, 341/77, 341/143, 375/242, 375/243, 375/245, 375/254, 395/1, 395/2.11, 395/2.91, 395/2.94, 395/2.95, 395/10-11, 395/20-23, 395/601, 395/606, 395/793, 395/804
• US Class Current: 382/232
• CPC Class Codes: G06T 9/005 Statistical coding, e.g. Huffman, run length coding
• CPC Class Codes: H04N 19/423 characterised by memory arrangements H04N19/433 takes precedence

Abstract

A method for compressing images uses artificial intelligence and neural-network-based techniques to convert digital image data into symbolic data. The symbolic data is then further compressed using, for example, run-length-limited coding. The result is compressed data that represents the original image data at a high compression ratio. Such high compression ratios are useful, for example, in medical diagnostic and high-definition television applications.

Claims

1. A method of compressing image data, comprising the steps of:

scanning an image to produce digital data corresponding to said image;

using artificial intelligence symbolics and neural network pattern recognition techniques, recognizing patterns in said digital data and simultaneously creating a look-up symbol table and symbolically-encoded data;

compressing said symbolically-encoded data to form compressed data;

processing said compressed data;

decompressing said compressed data to form decompressed symbolic data;

using said look-up symbol table, recovering digital image data from said decompressed symbolic data; and

providing said recovered digital image data to a display device.

2. A method as claimed in claim 1, wherein said step of processing comprises storing said compressed data.

3. A method as claimed in claim 1, where in the step of processing comprises the step of transmitting and receiving said compressed data.

4. A chip set on which an image compression method is implemented, said chip set comprising:

a processor means for executing said image compression method; and

memory means for use by said processor means, said memory means further comprising memory modules for storing program code implementing said image compression method, for use in executing the various steps of the method and for storing a color look-up table;

wherein said program code memory contains a module for creating a look-up table based on recognition of patterns in image data and for simultaneously symbolically encoding image data using artificial-intelligence-based and neural-network-based techniques and a module for further compressing said symbolically encoded image data.

5. A system for scanning, storing, retrieving and displaying images, comprising:

a scanning device;

an image processor for compressing digital image data generated by said scanning device and for decompressing compressed digital image data, comprising a chip set containing;

a processor means for executing an image compression method on said digital image data and for executing an inverse, decompression method on compressed digital image data; and

memory means for use by said processor means, said memory means further comprising memory modules for storing program code implementing said image compression method and its inverse decompression method, for use in executing the steps of the methods and for storing a color look-up table;

wherein said program code memory contains a module for creating a look-up table based on recognition of patterns in image data and for simultaneously symbolically encoding image data using artificial-intelligence-based and neural-network-based techniques;

a module for further compressing said symbolically encoded image data;

a module for decompressing compressed symbolically encoded image data; and

a module for decoding the decompressed symbolically encoded image data, based on said look-up table;

means for storing and retrieving compressed digital image data; and

means for displaying decompressed image data as a reconstructed image.

6. A system according to claim 5, wherein said neural-network-based techniques include the use of a multi-layer backpropagation-type neural network.

7. A method of compressing image data comprising the steps of:

scanning an image to produce digital data corresponding to the image;

creating a look up table using the image data and simultaneously translating a binary stream of the image data into encoded symbolic data, the encoded symbolic data having a known standard symbolic frequency distribution and bit code set;

compressing the encoded symbolic data to compressed symbolic data;

decompressing the compressed encoded symbolic data to form decompressed symbolic data;

converting the decompressed symbolic data into digital image data by associating the decompressed symbolic data with the look up table; and

providing the recovered digital image data to a display device.