Image data processing apparatus and image data processing method

• US 6,381,365 B2
• Filed: 08/20/1998
• Issued: 04/30/2002
• Est. Priority Date: 08/22/1997
• Status: Expired due to Fees

• Current Assignee: Minolta Co., Ltd. (Konica Minolta Inc.)
• Original Assignee: Minolta Co., Ltd. (Konica Minolta Inc.)
• Inventors: Murakawa, Akira
• Primary Examiner(s): Au, Amelia M.
• Assistant Examiner(s): Dastouri, Mehrdad
• Application Number:
  US09/137,234
• Publication Number:
  US 20010046321A1
• Time in Patent Office: 1,349 Days
• Field of Search: 382/190, 382/181, 382/209, 382/217, 382/218, 382/173, 382/170, 382/100, 358/521, 358/522
• US Class Current: 382/190
• CPC Class Codes: G06F 16/5862 using texture
• CPC Class Codes: G06T 7/49 based on structural texture description, e.g. using primitives or placement rules

Abstract

An image data processing apparatus for extracting a basic texture pattern as an identifiable feature of an image so that accurate image search can be performed by comparing images using the extracted basic texture patterns. The image data processing apparatus inputs image, detects textural periodicity in the image, extracts the basic texture pattern based on the calculated periodicity, normalizes the extracted basic pattern and shifts the normalized basic pattern. The apparatus stores the normalized and shifted basic pattern to an image database with the image from which the basic pattern was extracted.

Claims

1. An image data processing apparatus for processing image data of an image containing texture, said image data processing apparatus comprising:

a texture extraction controller for obtaining feature information relating to said texture from said image data;

a memory for storing image data and the feature information extracted by the texture extraction controller from the image data, said feature information being stored as pattern information and a lowest value of inertia of a gray level co-occurrence matrix; and

a comparison controller for comparing images based on the feature information stored to said memory.

2. The image data processing apparatus according to claim 1, wherein said feature information includes a basic pattern of a texture in the image, and arrangement information relating to a position of the basic pattern in the texture.

3. The image data processing apparatus according to claim 2, wherein said basic pattern is detected based on periodicity in two directions in the image, and said arrangement information includes the period in these two directions.

4. The image data processing apparatus according to claim 2, wherein the basic pattern is extracted such that significant pixels constituting the basic pattern are concentrated in a center part of the image.

5. The image data processing apparatus according to claim 1,

further comprising an input means for selecting an image to be used as a search key,

wherein said comparison controller searches for image data similar to the image selected by said input means by sequentially comparing the feature information for said selected image by said input means with the feature information of one or more images stored to said memory, based on a similarity between both feature information.

6. An image data processing apparatus for processing image data of an image containing a texture formed by a basic pattern repeating at a specific period, said image data processing apparatus comprising:

a detection controller for detecting feature information relating to the basic pattern of the texture and to arrangement of the basic pattern based on the image data;

a memory for storing image data and the feature information detected from the image data by said detection controller, said feature information being stored as pattern information and a lowest value of inertia of a gray level co-occurrence matrix; and

a comparison controller for comparing images based on the feature information stored to said memory.

7. The image data processing apparatus according to claim 6, wherein said detection controller detects periodicity in the image in two directions, and the arrangement information relates to the detected period and direction.

8. The image data processing apparatus according to claim 7, wherein the detection controller detects a basic pattern based on the detected periodicity in two directions.

9. The image data processing apparatus according to claim 6, wherein the basic pattern is detected such that significant pixels constituting the basic pattern are concentrated in a center of the image.

10. The image data processing apparatus according to claim 6,

further comprising an input means for selecting an image to be used as a search key,

wherein said comparison controller searches for image data similar to the image selected by said input means by sequentially comparing feature information for said selected image by said input means with the feature information of one or more images stored to said memory, based on a similarity between both feature information.

11. An image data processing apparatus for processing image data of an image containing a texture formed by a basic pattern repeating at a specific period, said image data processing apparatus comprising:

an image database for storing image data for a plurality of images and feature information comprising a basic pattern and period of a texture in the image, said feature information being stored as pattern information and a lowest value of inertia of a gray level co-occurrence matrix;

an input means for selecting an image to be used as a search key for seeking a desired image data from the image database; and

a controller for determining a similarity between the image selected by said input means and the image stored in said image database by sequentially comparing the feature information for the image selected by said input means with the feature information stored in the image database, and seeking a desired image from said image database based on said similarity.

12. The image data processing apparatus according to claim 11, wherein the period of the feature information comprises periods in two directions in the image.

13. The image data processing apparatus according to claim 11, further comprising:

an image data generator for generating image data to be stored to said image database; and

a database entry controller for extracting feature information from image data generated by said image data generator, and registering said extracted feature information with the image data in said image database.

14. A computer program product in a memory, said computer program product comprising:

a period detection control for detecting a period at which a basic pattern repeats in an image band of a defined width within an image containing a texture formed by a basic pattern that repeats at a specific period based on image data for said image; and

an extraction control for extracting the basic pattern of the texture in said image and the location of said basic pattern based on the period detected in the image band by said period detection step, wherein said period detection control and extraction control based upon a lowest value of inertia of a gray level co-occurrence matrix.

15. The computer program product in a memory according to claim 14, further comprising:

a storage control for storing image data and feature information detected by said extraction control to an image database, said image data and said feature information being stored with a defined relationship therebetween; and

a search control for seeking a desired image in said image database based on the feature information.

16. A computer program product in a memory, said computer program product comprising:

an input control for entering a search key used to seek a desired image from an image database, said image database storing image data for a plurality of images and image feature information related to said image, said feature information relating to a basic pattern of an image texture in the associated image and to a lowest value of inertia of a gray level co-occurrence matrix; and

a search control for seeking an image in said image database similar to the image selected as the search key by referencing the feature information.

17. The computer program product in a memory according to in claim 16, wherein said input control selects desired image and the feature information for said selected image is said search key.

18. An image data processing method comprising:

a detecting step for detecting a period of basic pattern repetition in an image band of a defined width of an image containing a texture pattern formed by a basic pattern repeating periodically; and

an extracting step for extracting the basic pattern and arrangement information thereof from said image data as feature information indicative of image features based on the period of basic pattern repetition detected in the image band by said detecting step, wherein said detecting step and said extracting step based upon a lowest value of inertia of a gray level co-occurrence matrix.

19. The image data processing method according to claim 18, further comprising:

a storing step for storing image data and feature information extracted by said extracting step with a relationship therebetween in an image database; and

a searching step for searching a desired image from said image database based on said feature information.

20. An image data processing method comprising:

a selecting step for selecting a search key used to seek a desired image from an image database; and

a searching step for seeking an image in said image database similar to the image selected as the search key by referencing the feature information,

said image database storing image data for a plurality of images and feature information for said images, said feature information being stored as pattern information and a lowest value of inertia of a gray level co-occurrence matrix, said feature information relating to a basic pattern of a texture in the associated image.

21. The image data processing method according to claim 20, wherein the feature information for an image selected by said selecting step is the search key.

22. An image data processing apparatus for processing image data for an image, said image having a matrix pattern consisting of a plurality of pixels and a texture pattern, said image data processing apparatus comprising

an extraction controller for extracting feature information indicative of a texture using pixel data for pixels located only within a particular area of the image, said particular area being a scanning band oriented in a predetermined direction through the image, the scanning band being a line or a band of a predetermined width, wherein said feature information based upon a lowest value of inertia of a gray level co-occurrence matrix.

23. The image data processing apparatus according to claim 22, wherein said extraction controller comprises:

a calculation means for calculating a gray level co-occurrence matrix for an image texture using pixel data for pixels located in a scanning band oriented in a predetermined direction through the image, the scanning band being a line or a band of a specific width; and

an extraction means for detecting image periodicity based on calculated results from said calculation means, and extracting said detected periodicity as an image feature.

24. The image data processing apparatus according to claim 22,

further comprising a digitizing means for digitizing gradation data of multiple gradation levels,

wherein the extraction means processes image data digitized by said digitizing means.

25. The image data processing apparatus according to claim 24, wherein said digitizing means digitizes using a value offset from a median value of a gradation distribution of pixels in said image data.

26. An image data processing method for processing image data for an image, said image having a matrix pattern consisting of a plurality of pixels and a texture pattern, said image data processing apparatus comprising

an extracting step for extracting feature information indicative of a texture using pixel data for pixels located only within a particular area of the image, the particular area being a scanning band oriented in a predetermined direction through the image, the scanning band being a line or a band of a predetermined width, wherein said feature information based upon a lowest value of inertia of a gray level co-occurrence matrix.

27. The image data processing method according to claim 26, wherein said extracting step comprises

a calculating step for calculating a gray level co-occurrence matrix for an image texture using pixel data for pixels located in a scanning band oriented in a predetermined direction through the image, said scanning band being a line or a band of a predetermined width;

and

a step for detecting image periodicity based on calculated results from said calculating step, and extracting the detected periodicity as a feature of image.

28. The image data processing method according to claim 26, further comprising a digitizing step for digitizing gradation data of multiple gradation levels, wherein said extracting step processes image data digitized by said digitizing step.

29. The image data processing method according to claim 28, wherein said digitizing step digitizes using a value offset from a median value of a gradation distribution of image data pixels.

30. A computer program product in a memory for processing image data for an image, said image consisting of a plurality of pixels arranged in a matrix pattern and having a texture pattern,

said computer program product comprising an extraction control for extracting feature information indicative of an image texture using pixel data for pixels located only within a particular area of the image, the particular area being a scanning band oriented in a predetermined direction through the image, the scanning band being a line or a band of a predetermined width, wherein said feature information based upon a lowest value of inertia of a gray level co-occurrence matrix.

31. The computer program product according to claim 30, wherein the extraction control further comprises:

a calculation control for calculating a gray level co-occurrence matrix for an image texture using pixel data for pixels located in a scanning band oriented in a predetermined direction through the image, the scanning band being a line or a band of a predetermined width; and

a extraction control for detecting image periodicity based on calculated results from said calculation control, and extracting the detected periodicity as an image feature.

32. The computer program product according to claim 30, further comprising a digitizing control for digitizing gradation data of multiple gradation levels, wherein said extraction control processes image data digitized by said digitizing control.

33. The computer program product according to claim 32, wherein said digitizing control digitizes using a value offset from a median value of a gradation distribution of image data pixels.