Cut mask preparation method and apparatus

• US 5,454,050 A
• Filed: 12/13/1994
• Issued: 09/26/1995
• Est. Priority Date: 12/20/1989
• Status: Expired due to Fees

• Current Assignee: Dai Nippon Printing Company Limited
• Original Assignee: Dai Nippon Printing Company Limited
• Inventors: Tanaka, Kazuhiko, Nakabayashi, Natsuko, Kobayashi, Hideaki
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US08/355,320
• Time in Patent Office: 287 Days
• Field of Search: 382/16, 382/17, 382/18, 382/22, 382/33, 382/50, 382/51, 382/52, 358/515, 358/517, 358/464
• US Class Current: 382/165
• CPC Class Codes: G03F 1/90 prepared by montage processes
• CPC Class Codes: G03F 3/06 by masking
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/136 involving thresholding
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation
• CPC Class Codes: H04N 1/3872 Repositioning or masking
• CPC Class Codes: H04N 1/46 Colour picture communication systems

Abstract

In a cut mask preparation method of displayed image on a screen of display means to prepare a cutting mask, an area lying across a real region and a background region is designated as a training area, for picture elements in which area a YMC three plate histogram is prepared to determine a threshold through discrimination analysis. A plate among the three plates, which has a maximum estimating function used upon the threshold determination, is used and binary-coded with said threshold to prepare a mask for a picture element included in one between the binary-coded values.

Claims

1. A cutting mask preparation method for preparing a cutting mask of color image data displayed on a monitor screen, said method comprising the steps of:

a) designating at least one training area for the displayed color image;

b) statistically analyzing said color image data in said at least one training area, and clustering a color distribution generated from said color image data in said at least one training area into a plurality of clusters based on the analysis so as to divide said color distribution into said plurality of clusters, said color distribution existent in a color space defined by color axes, each of said plurality of clusters defined by said color axes, said plurality of clusters formed by examining said color distribution along a selected one of said color axes to form intermediate clusters and repeatedly examining intermediate clusters along selected ones of said color axes to further form additional intermediate clusters until a classification termination condition is satisfied;

c) assigning a representative color to each one of the divided clusters, and designating at least one of the plurality of clusters to a background region, referring to said representative color;

d) designating a cutting area to said displayed color image data; and

e) binary-coding said displayed color image data within said designated cutting area to prepare the cutting mask by classifying pixel data within the designated cutting area as belonging to said at least one of said plurality of clusters designated to the background region.

2. The cutting mask preparation method as set forth in claim 1, wherein said statistical analysis and clustering are conducted by performing histogram discriminant analysis on each axis of the color distribution with respect to said color image data in said at least one training area, and said color image data is clustered along the axis of said plurality of axes having the greatest degree of separation of data of said color image.

3. The cutting mask preparation method as set forth in claim 1, wherein the pixel data within the designated cutting area is determined based on an estimating function prepared for each respective cluster, said estimating function indicating a possibility that an arbitrary color belongs to said respective cluster.