Color image region management system and color image region management method and color image region retrieval method

• US 5,509,111 A
• Filed: 05/22/1995
• Issued: 04/16/1996
• Est. Priority Date: 04/27/1992
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Sato, Machiko, Takahashi, Hiroyasu, Hong, Jung-Kook
• Primary Examiner(s): Herndon, Heather R.
• Assistant Examiner(s): VO, CLIFF N
• Application Number:
  US08/446,273
• Time in Patent Office: 330 Days
• Field of Search: 395/118, 395/131, 395/161, 395/132
• US Class Current: 345/591
• CPC Class Codes: G06T 11/001 Texturing; Colouring; Generation of texture or colour inpainting G06T5/77

Abstract

A color image region management system suitable for retrieving a region having a specified color. Color class is determined for each picture element of image data in which the color data of each picture element is expressed by a trichromatic system. In the trichromatic system, it is determined to which color class the color of the picture element belongs of a plurality of color classes divided on the basis of a perceptual color system. There is a color index memory for constructing tree structure data reaching a root node with blocks each including at least one picture element as a leaf node, and calculating, for each node in the tree structure data and for each color class in regards to the picture elements included in the node, the frequency of picture elements having a color belonging to the color class. The calculated frequency of the picture elements is stored as a color index for each color class.

Claims

1. A color image region management system comprising:

color class determination means for determining, for each picture element of image data in which color data of each picture element is expressed by a trichromatic system, to which color class the color of the picture element belongs of a plurality of color classes divided on the basis of a perceptual color system, and

color index memory means for constructing tree structure data reaching from leaf nodes to a root node, each leaf node including at least one picture element, and calculating, for each node in said tree structure data and for each color class in regards to the picture elements included in the node, a frequency of picture elements which have a color belonging to the color class, and storing in said each node the calculated frequency of picture elements as a color index for each color class.

2. A color image region management system as set forth in claim 1 wherein said frequency of picture elements is calculated for each node in said tree structure data and for each color class in regards to the picture elements included in the node, as a binary number which is one logical value if a picture element having a color belonging to the color class exists, and which is the other logical value if no picture element having a color belonging to the color class exists.

3. A color image region management system as set forth in claim 1 wherein said frequency of picture elements is calculated for each node in said tree structure data and for each color class in regards to the picture elements included in the node, as a count value of the number of picture elements which have a color belonging to the color class.

4. A color image region management system as set forth in claim 1 wherein, if the node in said tree structure data is a leaf node, said frequency of picture elements is calculated, for each color class in regards to the picture elements included in the node, as a binary number which is one logical value if a picture element having a color belonging to the color class exists, and which is the other logical value if no picture element having a color belonging to the color class exists, and if the node in said tree structure data is not a leaf node, said frequency of picture elements is calculated, for each color class in regards to the picture elements included in the node, as a count value of the number of the picture elements which have a color belonging to the color class.

5. A color image region management system as set forth in claim 3 wherein said count value of picture elements is quantized.

6. A color image region management system as set forth in claim 1 wherein at least two picture elements are included in each leaf node.

7. A color image region management system as set forth in claim 1 wherein only one picture element is included in each leaf node.

8. A color image region management system as set forth in claim 1 wherein said color class determination means comprises conversion means for converting color data of each picture element of image data from the trichromatic system to the perceptual color system, and means for dividing a color space based on said perceptual color system into a plurality of color classes and determining to which color class belongs the color converted to the perceptual color system by said conversion means.

9. A color image region management system as set forth in claim 8 wherein said color class determination means determines the color of a picture element as an achromatic color if the saturation of the color of the picture element is smaller than a predetermined value.

10. A color image region management method comprising the steps of:

converting color data of each picture element of image data from a trichromatic system to a perceptual color system,

dividing a color space based on the perceptual color system into a plurality of color classes, determining which color class of said plurality of color classes the color of a picture element belongs to, and

constructing tree structure data reaching from leaf nodes to a root node, each leaf node including at least one picture element, and for each node in said tree structure data and for each color class in regards to the picture elements included in the node, calculating a frequency of picture elements which have a color belonging to the color class, and storing in said each node the calculated frequency as a color index for each color class.

12. A color image region retrieval method for retrieving a color image region by the use of the tree structure data constructed by the color image region management method as set forth in claim 10 comprising the steps of:

determining the color classes to be retrieved on the basis of a specified color,

searching the tree structure data for the determined color classes,

in each node other than the leaf node, stopping the search of the nodes lower than itself if a frequency of the picture elements belonging to the determined color classes is zero, and otherwise continuing the search, and

in the leaf node, recording the leaf node as the region of the color to be detected if a frequency of the picture elements of the determined color classes is larger than a predetermined value.

13. A color image region retrieval method as set forth in claim 12 wherein said color classes to be retrieved are the color class to which the specified color belongs and color classes perceptually close to the color class to which the specified color belongs.

11. A color image region management system comprising:

an image input device for fetching an original image as image data expressed by a trichromatic system,

storage means for storing the image data expressed by the trichromatic system which is output from said image input device,

color class determination means for determining, for each picture element of the image data in which color data of each picture element is expressed by the trichromatic system, to which color class the color of the picture element belongs of a plurality of color classes divided on the basis of a perceptual color system, and

color index memory means for constructing tree structure data reaching from leaf nodes to a root node, each leaf node including at least one picture element, and for each node in said tree structure data and for each color class in regards to the picture elements included in the node, calculating a frequency of picture elements which have a color belonging to the color class, and storing in said each node the frequency as a color index for each color class.