Object boundaries identified in a raster image by a user selecting positions on the raster image and then using cost functions to predict likelihood of pixels near the position being on a boundary path

Claim 2

A computer-implemented method for identifying a boundary path in a displayed raster image, comprising: sampling user input to define cursor positions in a displayed raster image, wherein sequential cursor positions form a cursor path; defining one or more cost functions each a measure of a likelihood that a pixel falls on the boundary path; for each cursor position, using one of the one or more cost functions to evaluate pixels within a specified range of the cursor position to determine a cost associated with each pixel; and creating a boundary path of pixels based on the cursor path, wherein each boundary pixel corresponding to a cursor position in the cursor path has a minimum cost.

Claim 3

The method of claim 1, wherein user input is sampled periodically and boundary paths are created for more than one sampling.

Claim 4

The method of claim 1, further comprising: identifying a boundary start pixel as a start point, wherein boundary paths begin at the boundary start pixel; determining a boundary path segment of pixels from the boundary start pixel to an intermediate pixel, wherein the boundary path segment pixels are common to sequentially created boundary paths; saving the boundary path segment; setting the intermediate pixel as an updated boundary start pixel; and subsequently determining boundary paths beginning at the updated boundary start pixel.

Claim 5

The method of claim 1, further comprising displaying the boundary path.

Claim 6

The method of claim 4, further comprising editing the boundary path in response to user input.

Claim 7

The method of claim 1, wherein the step of creating a boundary path includes selecting a pixel from the pixels in the specified range having a minimum cost as a boundary pixel.

Claim 8

The method of claim 1, wherein a unique cost function is defined for each cursor position.

Claim 9

The method of claim 1 wherein the cost reflects a likelihood that a particular pixel falls on the boundary path.

Claim 10

The method of claim 1, wherein the cost is based on both the distance from a cursor position and a pixel value.

Claim 11

The method of claim 9, wherein a pixel value reflects whether an individual pixel has a boundary property.

Claim 12

The method of claim 10, wherein the boundary properties are selected from the group of Laplacian zero-crossings and gradient magnitude and direction.

Claim 13

The method of claim 9, wherein each cost function includes a tunnel width function defining a range around a cursor position within which the boundary pixel must lie.

Claim 14

The method of claim 12, wherein the tunnel width function is a variable function that assigns weights to pixels within the range based on their respective distance from the cursor position.

Claim 15

The method of claim 13, wherein the tunnel width function is determined based on user input.

Claim 16

The method of claim 12, wherein the range defines an area of the raster image based on the limits defined by the tunnel width function corresponding to a cursor position in the cursor path, and wherein the step of using a cost function further comprises: determining pixel values of pixels within the defined area of the raster image; applying a cost function to the pixel values of the pixels within the defined area of the raster image; and determining the boundary path based on the applied cost function.

Claim 17

The method of claim 15, further comprising determining pixel distances of pixels within the defined are of the raster image, wherein applying the cost function further comprises applying the cost function to the pixel distances of pixels within the defined area of the image.

Claim 18

A memory device storing computer-readable instructions for aiding a computer to identify a boundary path in a displayed raster image, comprising instructions for: sampling user input to define cursor positions in a displayed raster image, wherein sequential cursor positions form a cursor path; defining one or more cost functions each a measure of a likelihood that a pixel falls on the boundary path; for each cursor position, using one of the one or more cost functions to evaluate pixels within a specified range of the cursor position to determine a cost associated with each pixel and creating a boundary path of pixels based on the cursor path, wherein each boundary pixel corresponding to a cursor position in the cursor path has a minimum cost.

Claim 19

A memory device storing computer-readable instructions for aiding a computer to identify a boundary path in a displayed raster image, comprising instructions for: sampling user input to define cursor positions in a displayed raster image, wherein sequential cursor positions form a cursor path; defining one or more cost functions each a measure of a likelihood that a pixel falls on the boundary path; for each cursor position, using one of the one or more cost functions to evaluate pixels within a specified range of the cursor position to determine a cost associated with each pixel and creating a boundary path of pixels based on the cursor path, wherein each boundary pixel corresponding to a cursor position in the cursor path has a minimum cost.

Claim 20

A computer system for identifying a boundary path in a displayed raster image, comprising: a display for displaying a raster image; a user input device for receiving user input; and a processor for: sampling user input to define cursor positions in the displayed raster image, wherein sequential cursor positions form a cursor path; defining one or more cost functions each a measure of a likelihood that a pixel falls on the boundary path; for each cursor position, using one of the one or more cost functions to evaluate pixels within a specified range of the cursor position to determine a cost associated with each pixel; and creating a boundary path of pixels based on the cursor path, wherein each boundary pixel corresponding to a cursor position in the cursor path has a minimum cost.

Claim 21

A computer system for identifying a boundary path in a displayed raster image, comprising: a display for displaying a raster image; a user input device for receiving user input; and a processor for: sampling user input to define cursor positions in the displayed raster image, wherein sequential cursor positions form a cursor path; defining one or more cost functions each a measure of a likelihood that a pixel falls on the boundary path; for each cursor position, using one of the one or more cost functions to evaluate pixels within a specified range of the cursor position to determine a cost associated with each pixel; and creating a boundary path of pixels based on the cursor path, wherein each boundary pixel corresponding to a cursor position in the cursor path has a minimum cost.