Event clustering of images using foreground/background segmentation

Claim 2

An event clustering method for clustering images comprising similar events into same event image groups from a group of images, said event clustering method comprising the steps of: (a) segmenting each image within the group of images into a plurality of image regions, said image regions comprising at least one a foreground and background segmentation; (b) extracting at least one feature from each of a plurality of successive images from the group of images to form a feature set, each of said at least one feature deriving from said at least one foreground and background segmentation, said features comprising at least one of luminosity, color, position and size of said plurality of image regions; (c) computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set; (d) measuring a plurality of image distances between successive images as a function of said plurality of similarity measures; and (e) delimiting a set of event clusters as a function of said plurality of image distances, whereby each of said set of event clusters comprises a plurality of same event image groups, said same event image groups comprising images pertaining to a set of same events.

Claim 3

The method as claimed in claim 1 wherein the step (c) further comprises the step of computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set generating a distance measure that indicates the similarity or dissimilarity between said image regions.

Claim 4

The method as claimed in claim 1 wherein if a predetermined number of said image regions formed in step (a) are each less than a predetermined size, then a fixed image region is generated for the foreground within said foreground and background segmentation.

Claim 5

The method as claimed in claim 1 wherein the group of images are arranged in a chronological order and step (c) further comprises the step of computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set or estimating and comparing the similarity of said image regions comprising said foreground and background segmentation in every other image in the group of images and step (d) further comprises the step of measuring a plurality of image distances according to the total similarity between successive images and every other image.

Claim 6

The method as claimed in claim 1 wherein the group of images are arranged in a chronological order and step (c) further comprises the step of computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set estimating and comparing the similarity of said image regions comprising said foreground and background segmentation in every other two images in the group of images and step (d) comprises the step of measuring a plurality of image distances according to the total similarity between successive images and every other two images.

Claim 7

A computer storage medium having instructions stored therein for causing a computer for perform the method of claim 1 .

Claim 8

An event clustering method using foreground and background segmentation for clustering images from a group of images into same event image groups, said event clustering method including the steps of: (a) dividing each image into a plurality of blocks, thereby providing block-based images; (b) utilizing a block-by-block comparison to segment each block-based image into a plurality of regions comprising at least foreground and background segmentation; (c) extracting at least one feature from each of a plurality of successive images from the group of images to form a feature set, each of said at least one feature deriving from said at least one the foreground and background segmentation, said features comprising at least one of luminosity, color, position and size of the regions; (d) computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set, thereby leading to a plurality of image distance measures as a function of said plurality of similarity measures between successive images; and (e) delimiting a set of event clusters as a function of said plurality of image distances, whereby each of said set of event clusters comprises a plurality of same event image groups, said same event image groups comprising images pertaining to a set of same events.

Claim 9

An event clustering method using foreground and background segmentation for clustering images from a group of images into same event image groups, said event clustering method including the steps of: (a) dividing each image into a plurality of blocks, thereby providing block-based images; (b) utilizing a block-by-block comparison to segment each block-based image into a plurality of regions comprising at least foreground and background segmentation; (c) extracting at least one feature from each of a plurality of successive images from the group of images to form a feature set, each of said at least one feature deriving from said at least one the foreground and background segmentation, said features comprising at least one of luminosity, color, position and size of the regions; (d) computing a plurality of similarity measures for said plurality of successive images as a function of the similarity of said at least one feature within said feature set, thereby leading to a plurality of image distance measures as a function of said plurality of similarity measures between successive images; and (e) delimiting a set of event clusters as a function of said plurality of image distances, whereby each of said set of event clusters comprises a plurality of same event image groups, said same event image groups comprising images pertaining to a set of same events.

Claim 10

The method as claimed in claim 7 wherein the block-by-block comparison in step (b) comprises extracting one or more of said features from the blocks, utilizing the features to compute the similarity of each block with respect to its neighboring blocks, forming regions from similar blocks and merging similar regions into a background and a foreground segmentation.

Claim 11

A computer storage medium having instructions stored therein for causing a computer for perform the method of claim 7 .

Claim 12

The method as claimed in claim 7 wherein if a predetermined number of regions formed in step (b) are each less than a predetermined size, then a fixed regions is generated for the foreground.

Claim 13

An event clustering method using foreground and background segmentation for clustering images from a group of images into similar events, said method including the steps of: (a) dividing each image into a plurality of blocks, thereby providing block-based images; (b) utilizing a block-by-block comparison to segment each block-based image into a plurality of regions, wherein a first combination of regions comprises a foreground segmentation and a second combination of regions comprises a background segmentation; (c) extracting one or more features from said plurality of regions comprising said foreground segmentation and background segmentation, said features comprising at least one of luminosity, color, position and size of the regions; (d) computing a similarity measure between each region of the combination comprising a foreground segmentation of one image in the group and each region comprising the foreground segmentation of another image in the group of images, and further computing a similarity measure between each region of the combination comprising the background segmentation of said one image in the group of images and each region comprising the background segmentation of said another image in the group of images; (e) computing a mean value measure of the total similarity between successive images based on the similarity of all regions included in the combinations comprising said foreground segmentation and said background segmentation, thereby providing a measure of image distance between said images; and (f) delimiting a set of event clusters as a function of said plurality of image distances, whereby each of said set of event clusters comprises a plurality of same event image groups, said same event image groups comprising images pertaining to a set of same events.

Claim 14

An event clustering method using foreground and background segmentation for clustering images from a group of images into similar events, said method including the steps of: (a) dividing each image into a plurality of blocks, thereby providing block-based images; (b) utilizing a block-by-block comparison to segment each block-based image into a plurality of regions, wherein a first combination of regions comprises a foreground segmentation and a second combination of regions comprises a background segmentation; (c) extracting one or more features from said plurality of regions comprising said foreground segmentation and background segmentation, said features comprising at least one of luminosity, color, position and size of the regions; (d) computing a similarity measure between each region of the combination comprising a foreground segmentation of one image in the group and each region comprising the foreground segmentation of another image in the group of images, and further computing a similarity measure between each region of the combination comprising the background segmentation of said one image in the group of images and each region comprising the background segmentation of said another image in the group of images; (e) computing a mean value measure of the total similarity between successive images based on the similarity of all regions included in the combinations comprising said foreground segmentation and said background segmentation, thereby providing a measure of image distance between said images; and (f) delimiting a set of event clusters as a function of said plurality of image distances, whereby each of said set of event clusters comprises a plurality of same event image groups, said same event image groups comprising images pertaining to a set of same events.

Claim 15

The method as claimed in claim 11 wherein said (d) further comprises the step of computing a component to account for the relative sizes of the regions.

Claim 16

A computer storage medium having instructions stored therein for causing a computer for perform the method of claim 11 .

Claim 19

The method as claimed in claim 14 wherein said low-level features further comprise at least one of luminosity, color, position and size of said regions.

Claim 22

The system as claimed in claim 16 wherein the group of images are arranged in a chronological order and said third module further utilizes the features to estimate and compare the similarity of regions comprising foreground and background segmentation in every other image in the group of images and computes a measure of the total similarity between every other image, thereby providing image distance between successive images and every other image.

Claim 23

The system as claimed in claim 16 wherein the group of images are arranged in a chronological order and the third module further utilizes the features to estimate and compare the similarity of regions comprising foreground and background segmentation in every other two images in the group of images and computes a measure of the total similarity between every other two images, thereby providing image distance between successive images and every other two images.