Method of compressing a plurality of video images for efficiently storing, displaying and searching the plurality of video images

Claim 2

A single pass method of compressing a plurality of video images into a lesser number of key images, with the number of said lesser number of key images being specified, with each image characterized by a plurality of pixels, said method comprising the steps of: deriving at least one parameter for said plurality of video images; determining an image value for each video image for said one parameter; calculating a plurality of differences by comparing an image value associated with each video image to an image value associated with another video image; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference; and continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images.

Claim 3

The method of claim 1 wherein said another video image is immediately adjacent in time to said each video image.

Claim 4

The method of claim 1 wherein said deriving step derives a plurality of different parameters, each of which is derived by principal component analysis.

Claim 5

The method of claim 3 wherein said principal component analysis is performed locally with a different parameter for a different portion of each video image.

Claim 6

The method of claim 3 wherein said principal component analysis is performed globally with the same parameter used for the entirety of each video image.

Claim 7

The method of claim 5 wherein said determining step determines a plurality of image values for each video image, with one image value associated with one parameter.

Claim 8

The method of claim 6 wherein said calculating step calculates said difference in accordance with: ##EQU4## where k is an index to the plurality of parameters P is the total number of parameters I ik is the k parameter of image value I i I jk is the k parameter of image value I j W k is a weighing factor for parameter k.

Claim 9

The method of claim 7 wherein each of said weighing factor W k for all parameters k are the same.

Claim 10

The method of claim 8 wherein each of said weighing factor W k for all parameters k are not the same.

Claim 11

The method of claim 7 wherein I i and I j are image values associated with video images immediately adjacent in time.

Claim 12

The method of claim 7 wherein I i and I j are image values associated with video images not immediately adjacent in time.

Claim 13

The method of claim 1 wherein said one parameter is derived by discrete cosine transform.

Claim 14

The method of claim 1 wherein said one parameter is derived by Gabor filters.

Claim 15

The method of claim 1 wherein said one parameter is derived by wavelets.

Claim 16

The method of claim 1 wherein the number of key images is specified by a user.

Claim 17

The method of claim 1 wherein the number of key images is specified by a computer.

Claim 18

A single pass method of browsing by a user, a plurality of video images by compressing said plurality of video images into a lesser number of key images, with the number of said lesser number of key images being specified, with each image characterized by a plurality of pixels, said method comprising the steps of: deriving at least one parameter for said plurality of video images; determining an image value for each video image for said one parameter; calculating a plurality of differences by comparing an image value associated with each video image to an image value associated with another video image; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference; continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images; and displaying said video images not discarded.

Claim 19

A single pass method of browsing by a user, a plurality of video images by compressing said plurality of video images into a lesser number of key images, with the number of said lesser number of key images being specified, with each image characterized by a plurality of pixels, said method comprising the steps of: deriving at least one parameter for said plurality of video images; determining an image value for each video image for said one parameter; calculating a plurality of differences by comparing an image value associated with each video image to an image value associated with another video image; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference; continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images; and displaying said video images not discarded.

Claim 20

The method of claim 17 further comprising the step of: displaying the video images remaining after the discarding step to said user for browsing.

Claim 21

The method of claim 17 further comprising the step of: printing the video images remaining after the discarding step to said user for printing.

Claim 22

In a video database having a plurality of video scenes, each scene having a plurality of video images, each of said video images characterized by a plurality of pixels, a method for searching said video database to find a match to a first plurality of video images, said method comprising: compressing the video images of each video scene by: determining an image value for each video image for said one parameter for said video scene; calculating a plurality of differences by comparing an image value associated with each video image for said video scene to an image value associated with another video image of said video scene; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference of said video scene; continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images; storing said video images remaining and their image values; and comparing said image values associated with said first plurality of video images to said stored image values associated with each video scene to find a match between said first plurality of video images and one of said video scenes.

Claim 23

In a video database having a plurality of video scenes, each scene having a plurality of video images, each of said video images characterized by a plurality of pixels, a method for searching said video database to find a match to a first plurality of video images, said method comprising: compressing the video images of each video scene by: determining an image value for each video image for said one parameter for said video scene; calculating a plurality of differences by comparing an image value associated with each video image for said video scene to an image value associated with another video image of said video scene; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference of said video scene; continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images; storing said video images remaining and their image values; and comparing said image values associated with said first plurality of video images to said stored image values associated with each video scene to find a match between said first plurality of video images and one of said video scenes.

Claim 24

The method of claim 20 wherein said another video image for each scene is immediately adjacent in time to said each video image.

Claim 25

The method of claim 20 wherein said another video image for each scene is not adjacent in time to said each video image.

Claim 26

The method of claim 20 wherein said one parameter is derived by principal component analysis.

Claim 27

The method of claim 20 further comprising the steps of: compressing said first plurality of video images into a lesser number of key images, with the number of said lesser number of key images being specified by a user, by: deriving at least one parameter for said first plurality of video images; determining an image value for each video image for said one parameter; calculating a plurality of differences by comparing an image value associated with each video image to an image value associated with another video image; discarding the smallest difference, and one of the video images associated with an image value of said smallest difference; continuing with said step of discarding until the number of video images remaining equal to said lesser number of key images; and storing said video images remaining and their associated image values.

Claim 28

The method of claim 20 wherein said one parameter is derived by discrete cosine transform.

Claim 29

The method of claim 20 wherein said one parameter is derived by Gabor filters.

Claim 30

The method of claim 20 wherein said one parameter is derived by wavelets.