Using index keys extracted from JPEG-compressed images for image retrieval

• US 5,915,038 A
• Filed: 08/26/1996
• Issued: 06/22/1999
• Est. Priority Date: 08/26/1996
• Status: Expired due to Term

• Current Assignee: Phoenicia Innovations LLC Subsidiary of Pendrell Technologies LLC
• Original Assignee: Philips Electronics North America Corporation (Koninklijke Philips N.V.)
• Inventors: Abdel-Mottaleb, Mohamed Saeed, Shneier, Michael O.
• Primary Examiner(s): Coles, Edward L.
• Assistant Examiner(s): Lee, Cheukfan
• Application Number:
  US08/697,465
• Time in Patent Office: 1,030 Days
• Field of Search: 358/426, 358/262.1, 358/403, 358/261.4, 358/432, 358/435, 382/209, 382/224, 382/232, 382/235, 382/248, 382/250, 395/792, 395/605, 395/803, 395/773, 395/328, 395/207, 395/209, 395/222, 395/350, 395/615, 395/807
• US Class Current: 382/209
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06T 9/007 Transform coding, e.g. discrete cosine transform
• CPC Class Codes: H04N 1/41 Bandwidth or redundancy reduction by scanning H04N1/17 ; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00

Abstract

Index keys from JPEG encoded still images are extracted based on the gray-scale, luminance, and/or chrominance values. The index keys are stored in a database, along with corresponding location, and size information. The index key of a query image, also encoded in the JPEG format, is extracted. The index key of the image query is compared with the index keys stored in the meta database, with still images having index keys similar to the index key of the query image identified. The still images are then retrieved and displayed by selection of a user.

Claims

1. An apparatus comprising:

a source providing still images, said source comprising;

a JPEG server providing the still images encoded using a JPEG encoding standard,

a live image source providing the still images, and

a network source providing the still images encoded using an encoding standard;

an image information retrieval system, coupled to the source, extracting index keys from the still images and comparing an index key extracted from a query image to the index keys, said image information retrieval system comprising;

an index key extraction and archival section one of encoding and partially-re-encoding using the JPEG encoding standard the still images if the still images are not encoded using the JPEG encoding standard, and extracting the index keys from the still images if the still images are encoded using the JPEG encoding standard,

a database, coupled to the index key extraction and archival section, storing the index keys, along with identifying data of the respective still images corresponding to the index keys, and

a retrieval subsystem, coupled to the database, comparing the index key of a query image with the index keys stored in the database, and determining difference between the index key of the query image and each of the index keys stored in the database, said retrieval subsystem comprising;

a difference calculator calculating the difference,

an ordering unit, coupled to the difference calculator, ordering the index keys based on the difference, and

a fetch and display unit, coupled to the ordering unit, for retrieving still images corresponding to the index keys; and

a user interface, coupled to the image retrieval subsystem, displaying the still images.

2. A method for classifying image data, the method comprising executing the following steps in a digital data processing device:

recognizing an even number of windows that are subsets of the image;

applying a mathematical function to image data in each of the windows to yield a respective numerical value for each of the windows;

pairwise associating the windows to define at least one pair of windows;

comparing the respective numerical values within each of the at least one pair to yield a respective comparison value for each of the at least one pair;

forming an index key from the respective comparison value; and

embodying the index key in a medium readable by the digital data processing device.

3. A digital data processing device comprising:

a processor;

a device readable medium embodying the index key produced according to the method of claim 2.

4. The digital data processing device of claim 3 wherein the image data is a query image and the processor is adapted to retrieve images from a database using the index key as a query.

5. The digital data processing device of claim 3 wherein the image data is a data base image and the processor is adapted to match the index key to a query for retrieving the image data from the database.

6. The device of claim 5 wherein the medium also embodies the image data.

7. The device of claim 3 wherein the medium also embodies the image data.

8. The device of claim 3 wherein the processing device is adapted to accept user queries specifying a desired distance between a retrieved image and a query.

9. The device of claim 8 wherein the desired distance is specified by a Hamming measure.

10. The method of claim 2 wherein the image is a JPEG compressed image and each of the windows is a respective block resulting JPEG compression.

11. The method of claim 2 wherein the mathematical function comprises choosing a coefficient of the discrete cosine transform of each of the respective ones of the blocks.

12. The method of claim 2 wherein the coefficient is a DC coefficient.

13. The method of claim 2 wherein the mathematical function comprises an average of gray values in the window.

14. The method of claim 2 wherein the mathematical function comprises a variance of gray values in the window.

15. The method of claim 2 wherein the mathematical function comprises a number of edge points in the window.

16. A medium readable by a digital data processing device and embodying the index keys created by the method of claim 2.

17. The method of claim 2 wherein the image data comprises a stored database archive of images.

18. The method of claim 2 wherein the image data comprises a user-supplied query image.

19. The method of claim 2 wherein the image data comprises a stored database image.

20. The method of claim 2 wherein the processing step yields a plurality of index keys.