Image analysis – Image compression or coding – Interframe coding
Reexamination Certificate
1999-03-26
2001-07-10
Johns, Andrew W. (Department: 2621)
Image analysis
Image compression or coding
Interframe coding
Reexamination Certificate
active
06259817
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a method of compressing a plurality of video images (a video scene) such that the resultant compressed video images require less storage space than the compression methods of the prior art. In addition, the present invention efficiently displays the compressed video images, by displaying a reduced number of images or “key” images, thereby simplifying searching (by e.g. viewing by a user) of the contents of the video scene. The key images act similar to an index or a table of contents for the video scene.
BACKGROUND OF THE INVENTION
As is well known in the prior art, a video image, when digitized, requires a large amount of storage. A plurality of video images (as used hereinafter: “a video scene”), such as a movie, would require hundreds of megabytes or even gigabytes of storage, if not compressed.
Methods to compress a video scene are also well known in the prior art. One prior art method is to derive a parameter by principal component analysis for all the images of the video scene. Let us assume that each video image has N pixels. Thereafter, an image value based upon the parameter chosen by principal component analysis is determined for each of the video images. Thus, the resultant storage requirement, for one parameter, would be one image full of component values for that parameter (or N values) and one image value associated with each video image. If more than one parameter is derived for the video images, the total amount of storage required would be multiplied correspondingly.
However, even with this method, if the video scene has many video images, such as a movie, the storage required for a single value of the parameter associated with each of the video images would still be large. Since video images are displayed at a rather rapid rate, e.g., thirty times per second, a two hour video movie would require 216,000 video images (2×60×60×30) and at standard TV format would require 50 gigabyte of storage. Therefore, based on this method, for one parameter, there would still require N number of parameter component values, derived by principal component analysis, and 216,000 image values with one image value for each of the video images.
Furthermore, although storage requirement is reduced by this method, compared to the uncompressed video scene, to review or browse through the video images of this video scene would require the display of the entire collection of images, e.g. 216,000 images. To a user, viewing or searching through the entire collection of displayed images would be inefficient and cumbersome. Browsing or skimming through the entire collection of video images of the video scene would require a considerable mount of time and effort.
Another method of compression of the prior art is to choose some of the images from a video scene as reference frames. Subsequent or preceding images are compared to the reference frames, and the differences or changes are stored. Thus, data for only the reference frames and the changes thereto need to be stored.
However, in such prior art method, the reference frames are chosen based upon every nth frame. Thus, the choice of the reference frame is not optimized.
As the use of video images to capture information becomes more prevalent, a method must be found to efficiently store, display and search the plurality of video images, or the video scene. The collection of video images displayed must be categorized in a manner permitting rapid searching. Thus, not only is the compression of the video images important to minimize storage space, but the method must further facilitate the display and rapid searching of the compressed video images.
Hence, the present invention deals with the problems of video compression to minimize storage requirement as well as to permit a more efficient method of displaying and searching video images from a video scene.
SUMMARY OF THE INVENTION
In the present invention, a plurality of video images is compressed. Each image is characterized by a plurality of pixels. The method comprises the steps of deriving at least one parameter for the plurality of video images. The one parameter can be derived based upon principal component analysis. An image value is determined for each video image for that one parameter. An image value associated with each video image is compared to an image value associated with another video image to determine a difference. The difference is compared to a threshold, which can be fixed or adaptive. In the event the difference is below the threshold, only one of the image values is stored. The video images corresponding to the stored image values can be displayed and searched.
The aforementioned method can be used to compress a video database having a plurality of video images. One or more “key” video images are selected by the foregoing described method. All the other video images of the video database are compared to the “key” video images and the differences are stored. The “key” video images, by themselves, or the principal component analysis image and the parameter values determined, are also stored. Based upon this compression, all the video images of the video database can be faithfully reproduced.
In another method of the present invention, a video database has a plurality of database elements with each database element being a video scene. Each video scene is compressed based upon the foregoing disclosed compression method. The video database is searched to find a match with an inquiring video scene, also comprising a plurality of video images. The inquiring video scene may also be compressed by the foregoing disclosed compression method. The stored image values associated with the inquiring video scene are compared to the stored image values associated with each video scene from the video database to find a match between the inquiring video scene and one of the video scenes from the video database.
REFERENCES:
patent: 4217609 (1980-08-01), Hatori et al.
patent: 5157742 (1992-10-01), Niihara
patent: 5177796 (1993-01-01), Feig et al.
patent: 5226093 (1993-07-01), Iwase
patent: 5265180 (1993-11-01), Golin
patent: 5299019 (1994-03-01), Pack et al.
patent: 5305400 (1994-04-01), Butera
patent: 5809161 (1998-09-01), Auty et al.
patent: 5974235 (1999-10-01), Nunally et al.
Computer Graphics Animation of Talking Faces Based on Stochastic Modesl , N.M. Brooks et al., ISSIPNN '94, 1994 International Symposium on Speech, Image Processing and Neural Networks Proceedings, Apr. 1994, pp. 73-76.
Welsh, W.J. et al., “Facial Feature Image Coding Using Principal Components”, Electronics Letters, vol. 28, No. 22 Oct. 22, 1992, pp. 2066-2067.
Azarian Seyed H.
Gray Cary Ware & Freidenrich LLP
Interval Resarch Corporation
Johns Andrew W.
LandOfFree
Method of compressing a plurality of video images for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of compressing a plurality of video images for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of compressing a plurality of video images for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2441001