Method and apparatus for segmenting images

A method of segmenting an image into regions, the image comprising a plurality of pixels, said method comprising the steps of:

• (a) allocating one or more pixels as seeds for the regions;
• (b) generating a list of pixels that border the regions;
• (c) scanning only a subset of pixels in the list of pixels, the number of pixels in the subset being smaller than the number of pixels in the list;
• (d) determining a value, for each scanned pixel, indicative of the similarity of a property of the scanned pixel and the corresponding property of a region that the scanned pixel borders;
• (e) selecting a scanned pixel that is most similar in the property to the region that borders the scanned pixel, wherein, if a current scanned pixel has a value below a minimum threshold said scanning step terminates and the current scanned pixel is selected;
• (f) appending the selected pixel to the region bordered by the selected pixel;
• (g) calculating an updated property of the region as a function of the property of the region prior to said appending step and the property of the appended pixel;
• and
• (h) repeating steps (b) to (g) until the image is segmented.

A method of segmenting an image, the image comprising a plurality of pixels, said method comprising the steps of:

• allocating one or more pixels as seeds in the image;
• and
• growing regions of pixels from the seeds, wherein said growing step comprises the sub-steps of:
• generating a list of pixels that border the regions;
• scanning a number of the pixels of the list using a step size between scanned pixels that is a function of the length of the list;
• determining, for each scanned pixel, a value indicative of the similarity of the luminance of the scanned pixel and the corresponding luminance of a region that the scanned pixel borders;
• selecting a pixel that has a minimum value;
• appending the selected pixel to the region it borders;
• updating the corresponding luminance of the region that includes the appended pixel;
• and
• repeating the sub-steps of the growing step until there are no more pixels that border the regions.

A method of segmenting an image, the image comprising a plurality of pixels and the method comprising the steps of:

• allocating one or more pixels as seeds;
• growing regions from the seeds so as to segment the image into a number of regions;
• storing the segmented image in a queue;
• and
• performing the following sub-steps until the queue is empty:
• removing and selecting the segmented image or a previously divided rectangular area of the segmented image currently stored first in the queue as the current rectangular area;
• computing a value representative of the size of a dominant segmented region within the current area divided by the size of the current area;
• storing the co-ordinates of the current area, if the value is greater than a predetermined threshold, otherwise;
• dividing the current area into a plurality of rectangular areas and adding the plurality of rectangular areas to the queue;
• merging the divided rectangular areas which have a common dominant region and share a common edge;
• and
• outputting the vertices of the merged rectangular areas as a representation of the segmented image.

A method of segmenting an image, the image comprising a plurality of pixels, said method comprising the steps of:

• distributing seeds in areas of the image as a function of a property of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels homogeneous in the property, and wherein said distribution step comprises the sub-steps of:
• (a) dividing the image into a plurality of areas;
• (b) allocating for each divided area, one or more seeds as a function of the property within the divided area and the size of the divided area as compared to the size of the image;
• and
• (c) storing the locations of each allocated seed;
• growing regions from the seeds so as to segment the image into regions, wherein only a number of pixels that border the regions are considered, the number being smaller than a total number of pixels that border the regions, and the considered pixel that is most similar in a property to a region bordered by the considered pixel is appended to the region to form an expanded region and the property of the expanded region is updated;
• and
• repeating said growing step until no pixels bordering the regions are available.

A method of segmenting an image, the image comprising a plurality of pixels, said method comprising the steps of:

• (a) allocating pixels as seeds in areas of the image as a function of the luminance of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels of homogeneous luminance and wherein the seeds form growing regions;
• (b) generating a list of pixels that border the growing regions;
• (c) scanning a number of the pixels of the list of pixels using a step size between scanned pixels that is a function of the length of the list;
• (d) determining, for each scanned pixel, a value indicative of the similarity of the luminance of the scanned pixel and the corresponding luminance of a growing region that the scanned pixel borders;
• (e) selecting a pixel that has a minimum value;
• (f) appending the selected pixel to the growing region it borders;
• (g) updating the corresponding luminance of the region that includes the appended pixel;
• and
• (h) repeating the sub-steps (b) to (g) until there are no more pixels that border the growing regions.

Apparatus for segmenting an image, the image comprising a plurality of pixels, wherein said apparatus comprises:

• means for allocating one or more pixels as seeds;
• means for growing regions from the seeds so as to segment the image into a number of regions, wherein only a number of pixels that border the regions are considered, the number being smaller than a total number of pixels that border the regions, and wherein a variable step size is used to scan the bordering pixels, the step size being a function of the quantity of bordering pixels, and the considered pixel that is most similar in a property to a region bordered by the considered pixel is appended to the region to form an expanded region and the property of the expanded region is updated;
• and
• means for controlling to effect repeated application of said means for growing regions until no pixels bordering the regions are available.

Apparatus for segmenting an image, the image comprising a plurality of pixels, wherein said apparatus comprises:

• means for allocating one or more pixels as seeds in the image;
• and
• means for growing regions of pixels from the seeds, wherein said growing means comprises:
• means for generating a list of pixels that border the regions;
• means for scanning only a subset of the pixels of the list, the number of pixels in the subset being smaller that the number of pixels in the list;
• means for determining, for each scanned pixel, a value indicative of the similarity of a property of the scanned pixel and the corresponding property of a region that the scanned pixel borders;
• means for selecting a scanned pixel that is most similar in the property to the region that the scanned pixel borders, wherein, if a current scanned pixel has a value below a minimum threshold, said means for scanning stops scanning the subset of pixels and said means for selecting selects the current scanned pixel;
• means for appending the selected pixel to the region bordered by the selected pixel;
• means for calculating an updated property of the region as a function of the property of the region prior to the appending and the property of the appended pixel;
• and
• means for repeating the operations of said growing means until there are no more pixels that border the regions.

Apparatus for segmenting an image, the image comprising a plurality of pixels and the apparatus comprising:

• means for allocating one or more pixels as seeds;
• means for growing regions from the seeds so as to segment the image into a number of regions;
• means for storing the segmented image in a queue;
• means for removing and selecting, until the queue is empty, the segmented image or a previously divided rectangular area of the segmented image currently stored first in the queue as the current rectangular area;
• means for computing a value representative of the size of the dominant segmented region within the current area divided by the size of the current area;
• means for storing the co-ordinates of the current area, if the value is greater than a predetermined threshold;
• means for dividing the current area into a plurality of rectangular areas of the current area and adding the plurality of rectangular areas to the queue, if the value is less than or equal to the predetermined threshold;
• means for merging the divided rectangular areas which have a common dominant region and share a common edge;
• and
• means for outputting the vertices of the merged rectangular areas as a representation of the segmented image.

Apparatus for segmenting an image, the image comprising a plurality of pixels, wherein said apparatus comprises:

• means for distributing seeds in areas of the image as a function of a property of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels homogeneous in the property, and wherein said means for distributing seeds comprises:
• (a) means for dividing the image into a plurality of areas:
• (b) means for allocating, for each divided area, one or more seeds as a function of the property within the divided area and the size of the divided area as compared to the size of the image;
• and
• (c) means for storing the locations of each allocated seed;
• means for growing regions from the seeds so as to segment the image into regions, wherein only a number of pixels that border the regions are considered, the number being smaller than a total number of pixels that border said regions, and the considered pixel that is most similar in a property to a region bordered by the considered pixel is appended to the region to form an expanded region and the property of the expanded region is updated;
• and
• means for controlling to effect repeated application of said means for growing regions until no pixels bordering the regions are available.

Apparatus for segmenting an image, the image comprising a plurality of pixels, wherein said apparatus comprises:

• means for allocating pixels as seeds in areas of the image as a function of the luminance of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels of homogeneous luminance and wherein the seeds form growing regions;
• means for generating a list of pixels that border the growing regions;
• means for scanning a number of the pixels of the list of pixels using a step size between scanned pixels that is a function of the length of the list;
• means for determining, for each scanned pixel, a value indicative of the similarity of the luminance of the scanned pixel and the corresponding luminance of a growing region that the scanned pixel borders;
• means for selecting a pixel that has a minimum the value;
• means for appending the selected pixel to the growing region it borders;
• means for updating the corresponding luminance of the region that includes the appended pixel;
• and
• means for repeating the operations of the allocating means, generating means, scanning means, determining means, appending means, and updating means until there are no more pixels that border the growing regions.

A computer program product, including a computer readable medium having recorded thereon a computer program for segmenting an image, the image comprising a plurality of pixels, wherein said computer program product comprises:

• means for allocating one or more pixels as seeds;
• means for growing regions from the seeds so as to segment the image into regions, wherein only a number of pixels that border the regions are considered, the number being smaller that a total number of pixels that border the regions, and wherein a variable step size is used to scan the bordering pixels, the step size being a function of the quantity of bordering pixels, and the considered pixel that is most similar in a property to a region bordered by the considered pixels is appended to the region to form an expanded region and the property of the expanded region is updated;
• and
• means for controlling to effect repeated application of said means for growing regions until no pixels bordering the regions are available.

A computer program product, including a computer readable medium having recorded thereon a computer program for segmenting an image, the image comprising a plurality of pixels, wherein said computer program product comprises:

• means for allocating one or more pixels as seeds in the image;
• and
• means for growing regions of pixels from the seeds, wherein said growing means comprises:
• means for generating a list of pixels that border the regions;
• means for scanning only a subset of the pixels of the list, the number of pixels in the subset being smaller than the number of pixels in the list;
• means for determining, for each scanned pixel, a value indicative of the similarity of a property of the scanned pixel and the corresponding property of a region that the scanned pixel borders;
• means for selecting a scanned pixel that is most similar in the property to the region that the scanned pixel borders, wherein, if a current scanned pixel has a value below a minimum threshold, said means for scanning stops scanning the subset of pixels and said means for selecting selects the current scanned pixel;
• means for appending the selected pixel to the region bordered by the selected pixel;
• means for calculating an updated property of the region as a function of the property of the region prior to the appending and the property of the appended pixel;
• and
• means for repeating the operations of said growing means until there are no more pixels that border the regions.

A computer program product, including a computer readable medium having recorded thereon a computer program for segmenting an image, the image comprising a plurality of pixels and the computer program product comprising:

• means for allocating one or more pixels as seeds;
• means for growing regions from the seeds so as to segment the image into a number of regions;
• means for storing the segmented image in a queue;
• means for removing and selecting, until the queue is empty, the segmented image or a previously divided rectangular area of the segmented image currently stored first in the queue as the current rectangular area;
• means for computing a value representative of the size of the dominant segmented region within the current area divided by the size of the current area;
• means for storing the co-ordinates of the current area, if the value is greater than a predetermined threshold;
• means for dividing the current area into a plurality of rectangular areas of the current area and adding the plurality of rectangular areas to the queue, if the value is less than or equal to the predetermined threshold;
• means for merging the divided rectangular areas which have a common dominant region and share a common edge;
• and
• means for outputting the vertices of the merged rectangular areas as a representation of the segmented image.

A computer program product, including a computer readable medium having recorded thereon a computer program for segmenting an image, the image comprising a plurality of pixels, wherein said computer program product comprises:

• means for distributing seeds in areas of the image as a function of a property of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels homogeneous in the property, and wherein said means for distributing seeds comprises:
• (a) means for dividing the image into a plurality of areas;
• (b) means for allocating, for each divided area, one or more seeds as a function of the property within the divided area and the size of the divided area as compared to the size of the image;
• and
• (c) means for storing the locations of each allocated seed;
• means for growing regions from the seeds so as to segment the image regions, wherein only a number of pixels that border the regions are considered, the number being smaller than a total number of pixels that border the regions, and the considered pixel that is most similar in a property to a region bordered by the considered pixel is appended to the region to form an expanded region, and the property of the expanded region is updated;
• and
• means for controlling to effect repeated application of said means for growing regions is repeatedly applied until no pixels bordering the regions are available.

A computer program product, including a computer readable medium having recorded thereon a computer program for segmenting an image, the image comprising a plurality of pixels, wherein said computer program product comprises:

• means for allocating pixels as seeds in areas of the image as a function of the luminance of the pixels within those areas, wherein fewer seeds are allocated to those areas of the image having pixels of homogeneous luminance and wherein the seeds form growing regions;
• means for generating a list of pixels that border the growing regions;
• means for scanning a number of the pixels of the list of pixels using a step size between scanned pixels that is a function of the length of the list;
• means for determining, for each scanned pixel, a value indicative of the similarity of the luminance of the scanned pixel and the corresponding luminance of a growing region that the scanned pixel borders;
• means for selecting a pixel that has a minimum value;
• means for appending the selected pixel to the growing region it borders;
• means for updating the corresponding luminance of the region that includes the appended pixel;
• and
• means for repeating the operations of said allocating means, generating means, scanning means, determining means, appending means, and updating means until there are no more pixels that border the growing regions.