Image editing system and method having improved automatic object selection

• US 5,526,020 A
• Filed: 08/03/1993
• Issued: 06/11/1996
• Est. Priority Date: 10/07/1991
• Status: Expired due to Term

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Venable, Dennis L., Campanelli, Michael R.
• Primary Examiner(s): Hjerpe, Richard
• Assistant Examiner(s): Stoll, Kara Farnandez
• Application Number:
  US08/101,097
• Time in Patent Office: 1,043 Days
• Field of Search: 382/17, 382/18, 382/48, 382/50, 382/51, 382/52, 340/703, 358/22, 358/453, 345/145, 345/157, 345/150, 345/153, 395/155
• US Class Current: 382/165
• CPC Class Codes: G06F 3/04842 Selection of displayed objects or displayed text elements G06F3/0482 takes precedence

Abstract

An image editor stores color pixels for an image to be edited in an image pixel buffer and a lookup register bank. The stored image is displayed on a monitor. An operator uses a mouse to designate a screen path that identifies an object in the image, and pixel position data is detected for the screen path. A programmed computer determines in real time one or more criteria for a search controller on the basis of hue, saturation and luminance variables which are determined from the stored pixel color data for pixels along the screen path. The computer further searches the image pixels to identify the object and its boundaries in accordance with the search controller criteria, and applies color edit actions to the identified object within the image.

Claims

1. An image editor for an image processing system, said editor comprising:

means for storing pixel color data for an image to be edited;

means for displaying the stored image on a monitor screen;

means for designating a screen path which identifies an object and which provides position and color date for use in determining pixels corresponding to the object in the displayed image by differentiating the object from its nearby context;

means for detecting pixel position data corresponding to the screen path;

means for determining in real time one or more criteria for a search controller as a function of screen path position data and screen path color data;

means for searching image pixels to find those pixels diverging from the screen path and satisfying the criteria set by the search controller thereby determining a contiguous image area corresponding to the image area occupied by the selected object and thereby identifying the selected object relative to its context; and

means for applying a preselected color edit operation to the identified object within the image.

2. The image editor of claim 1 wherein said search criteria determining means includes:

means for determining one or more color data variables from the stored pixel color data for pixels along the screen path; and

means for determining said criteria for the search controller at least as a function of the determined color data variables.

3. The image editor of claim 1 wherein said storing means includes:

a color lookup register bank in which respective registers contain respective color values for respective colors in a preselected color set where each color value is defined by a set of values for preselected basic color components;

a pixel buffer in which the color of each image pixel is stored as a lookup register value; and

wherein said color edit applying means includes;

means for implementing color edit actions by modifying stored color values for lookup registers containing original colors changed by edit action; and

means for selecting image buffer pixels corresponding to the identified object for application of edit actions.

4. The image editor of claim 1 wherein:

the stored pixel color data are values for preselected basic colors;

said search criteria determining means includes means for determining, in respective multiple data channels, values for preselected multiple color descriptor variables from the detected basic color data values; and

the available search criteria, from which the one or more selected search criteria are determined, include respective criteria based on the respective color descriptor variable channels.

5. The image editor of claim 4 wherein said color descriptor variables are hue, saturation and luminance.

6. The image editor of claim 4 wherein the available search criteria include a user specification, and, if entered, a user specification selected as the search criteria.

7. The image editor of claim 6 wherein the available search criteria include a synthetic or solid color, and luminance only, and the selected criterion is the synthetic or solid color or luminance if so indicated by the color data for detected pixels.

8. The image editor of claim 4 wherein the selected search criterion is one of the color descriptor data channels if said one channel is the only color descriptor channel having data values for the detected pixels.

9. The image editor of claim 1 wherein said search criteria determining means includes means for statistically analyzing said stored pixel color data in accordance with predetermined statistical principles to select as said search criteria a portion or portions of said stored color pixel data that will most likely provide effective automatic identification of said object.

10. The image editor of claim 9 wherein:

the stored pixel color data are values for preselected basic colors; and

each statistically selected color data portion is a portion of the color data in one of said color descriptor variable channels.

11. The image editor of claim 10 wherein said color descriptor variables are hue, saturation and luminance.

12. The image editor of claim 9 wherein:

said statistical analyzing means generates a histogram that functionally relates frequency of color data to color data values; and

each selected color data portion is a node of said histogram.

13. The image editor of claim 10 wherein:

said statistical analyzing means generates a histogram that functionally relates frequency of color data to color data values; and

said histogram includes a component histogram for each color descriptor variable channel.

14. The image editor of claim 13 wherein the color descriptor channels are hue, saturation and luminance.

15. The image editor of claim 12 wherein said image searching means executes an image data search in accordance with each node and sums the results.

16. The image editor of claim 13 wherein:

said search criteria determining means includes means for sorting color descriptor channels by range of standard deviation and for excluding as unusable those channels having a standard deviation below a preselected threshold value; and

said search criteria determining means selects for use as search criteria those channels having a standard deviation above said threshold value.

17. The image editor of claim 16 wherein the color descriptor channels are hue, saturation and luminance.

18. The image editor of claim 17 wherein the selected search criteria is hue if no other channel is usable.

19. The image editor of claim 1 wherein:

means are provided for managing a window structure on said screen with said window structure including a window for said image and a menu window for predetermined selectors applicable to said image;

said selectors include an automatic object selector; and

said screen path designating means is employed to select said automatic object selector when it is desired to apply edit actions only to an object within said image.

20. The image editor of claim 19 wherein said screen designating means is a mouse.

21. The image editor of claim 1 wherein said pixel position detecting means gathers color pixel data for a pixel band of predetermined pixel width along said path.

22. An automatic object selector for an image processing system in which color data is stored for pixels of an image to be displayed on a screen for editing, said selector comprising:

means for detecting pixel position data corresponding to a path designated on said screen to identify an object occupying an area of the displayed image;

means for determining in real time one or more criteria for a search controller as a function of the stored pixel color data for pixels along the screen path;

means for searching the image pixels to find those pixels diverging from the screen path and satisfying the search controller criteria thereby determining a contiguous image area corresponding to the image area occupied by the selected object and thereby identifying the object and its boundary; and

means for applying color edit actions to the identified object within the image.

23. The automatic object selector of claim 22 wherein:

the stored pixel color data are values for preselected basic colors;

said search criteria determining means includes means for determining, in respective multiple data channels, values for preselected multiple color descriptor variables from the detected basic color data values; and

the available search criteria, from which the one or more selected search criteria are determined, include respective criteria based on the respective color descriptor variable channels.

24. The automatic object selector of claim 23 wherein said color descriptor variables are hue, saturation and luminance.

25. The automatic object selector of claim 22 wherein said search criteria determining means includes means for statistically analyzing said stored pixel color data in accordance with predetermined statistical principles to select as said search criteria a portion or portions of said stored color pixel data that will most likely provide effective automatic identification of said object.

26. The automatic object selector of claim 25 wherein:

said statistical analyzing means generates a histogram that functionally relates frequency of color data to color data values; and

each selected color data portion is a node of said histogram.

27. A process for editing images in an image processing system, the steps of said process comprising:

storing color data for pixels of an image to be edited;

displaying the stored image on a monitor;

designating a screen path that identifies an object in the image;

detecting pixel position data corresponding to the screen path;

determining pixel color data corresponding to the pixel position data;

determining in real time at least one criterion for controlling an object search as a function of the pixel color data for pixels along the screen path;

searching the image pixels to find those pixels diverging from the screen path and satisfying the search controller criteria thereby determining a contiguous image area corresponding to the area occupied by the selected object and thereby identifying the object and its boundary relative to its context; and

applying color edit actions to the identified object image within the image.

28. The process of claim 27 wherein:

the stored pixel color data are values for preselected basic colors; and

wherein said search criteria determining step includes;

selecting values for preselected multiple color descriptor variables from the detected basic color data values; and

using said color descriptor variable channels as available search criteria.

29. The process of claim 27 wherein said search criteria determining step includes statistically analyzing the stored pixel color data in accordance with predetermined statistical principles to select as the search criteria a portion or portions of the stored pixel color data that most likely will provide effective automatic identification of the object.

30. The process of claim 29 wherein said analyzing step includes generating a histogram that functionally relates frequency of color data to color data values with each color data portion being a node of said histogram.

31. An image editor for an image processing system, said editor comprising:

means for storing pixel color data for an image to be edited;

means for displaying the stored image on a monitor screen;

means for designating a screen path which identifies an object and which provides position and color data for use in determining pixels corresponding to the object in the displayed image by differentiating the object from its nearby context, the object occupying an area of the displayed image having multiple pixels with differently valued color variables;

means for detecting pixel position data corresponding to the screen path;

means for determining in real time one or more criteria for a search controller as a function of screen path position data and screen path color data;

means for searching image pixels to find those pixels diverging from the screen path and satisfying the criteria set by the search controller thereby determining a contiguous image area corresponding to the image area occupied by the selected object and thereby identifying the selected object relative to its context; and

means for applying a preselected color edit operation to the identified object within the image.