Image analysis – Image transformation or preprocessing
Reexamination Certificate
1999-06-21
2001-06-12
Mancuso, Joseph (Department: 2723)
Image analysis
Image transformation or preprocessing
C348S086000
Reexamination Certificate
active
06246803
ABSTRACT:
COPYRIGHT NOTICE AND AUTHORIZATION
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.
FIELD OF THE INVENTION
This invention relates to video stream comparison and distortion analysis to determine variations between two or more video streams and, more particularly, to real-time video stream comparison and distortion analysis utilizing frame-by-frame comparisons of baseline color and intensity moments with broadcast color and intensity moments to determine variations between two or more video streams.
BACKGROUND OF THE INVENTION
Television broadcasting has become prevalent in almost every part of the world. Broadcast signals are continuously transmitted and retransmitted to millions of homes around the world through cable, satellite, and other mediums. The majority of revenues required to operate these television networks are provided by advertisers, and advertisers generally desire some confirmation that the advertising time they paid for was actually transmitted to the viewers. In the broadcast industry, such confirmation is referred to as video reconciliation. Video authentication broadly refers to the process of verifying that a received signal matches identically that which has been broadcast with cable, satellite, or other wireless networks, and reconciliation is the authentication of small video segments such as commercials. Thus, reconciliation is performed by broadcasters to ensure that advertising commercials have been properly inserted for retransmission by affiliate stations and networks. Further, the quality of these transmissions is important to both paying cable and satellite customers and non-paying customers which simply receive a broadcast signal with an antenna and tuner.
To confirm/reconcile that advertisements have been aired and to check the quality of transmissions, two video streams can be compared to find distortions. Some broadcasters employ a laborious and costly approach utilizing paid personnel to reconcile advertising commercials by manually viewing recorded segments and relaying variations to the broadcaster. Disadvantageously, it can take as long as two months before the broadcaster receives a reconciliation report. Because it is not commercially desirable for individuals to continuously monitor and compare a broadcast video stream with a baseline video stream for commercial advertisement reconciliation and quality control, attempts have been made to recharacterize the video streams for comparison by a computer. Further, if two video streams are to be compared at a remote location, the baseline video stream must be recharacterized for transmission to the remote location by a transmission method more reliable and preferably faster than the broadcast signal carrying the broadcast video stream.
The present invention is an important and novel improvement over currently employed systems and methods, as well as processes and methods previously identified in publications and U.S. and foreign patents for recharacterizing and comparing two video streams. For example, as applied to video authentication, the primary advantage of the present invention is that it does not require the insertion of digital or analog “watermarks” into the video stream for tracking or detection. One such advertisement reconciliation system is discussed in U.S. Pat. No. 5,826,165 to Echeita, et al., and pertains to digital video (not analog National Television Standards Committee (NTSC) signals), such as those currently broadcast via direct broadcast satellite (DBS), into which “reconciliation data packets” are inserted into the digital video signal to identify commercials. The reliance upon digital video, video message format standards, and the added complexity of broadcast signal manipulation limits the scope and commercial viability of this approach.
Similarly, other video authentication techniques have been described in the literature, which rely on embedding visible or invisible watermarks within the video signal. One approach is to hide the watermark pattern in the least significant bit of each pixel in each frame. Another approach is to embed this information in the low frequency components of each frame such that the watermark is preserved after an image has been compressed or decompressed using traditional compression techniques such as JPEG (Joint Photographic Experts Group) or MPEG (Moving Pictures Experts Group). One commercially advertised system employing this technique is currently offered by Nielsen Media Research (New Media Services Division) under the trade name of “SIGMA”, which embeds an “invisible” code into the video stream for analysis at the signal's receiving end. This embedded code, as described in the literature and through the company's web site, includes encode number, date, and time stamp.
The fundamental difficulty with all water marking schemes is that certain video manipulations such as smoothing can accidentally (or intentionally) alter or destroy the watermark. The present invention avoids these difficulties and simplifies the video delivery process by eliminating the need for watermarks.
Video segment recognition is one application that is related to video distortion analysis, and is described in prior art. For example, U.S. Pat. No. 4,739,398 by Thomas, et al., describes a system, method, and apparatus for recognizing broadcast segments in analog video signals which relies on building a hash table of digital “signatures.” Each signature is based on the average luminance for randomly selected frames within a video segment, as well as selected locations within each analyzed frame. The main disadvantages to this method are: (1) two very different video segments can have the same hash table signature, (2) small changes to a single frame can result in very different signatures, and (3) comparing signatures provides no information about spatial, chromatic, or temporal differences between video segments. These factors preclude its use for video stream distortion analysis, as well as its intended application for video reconciliation.
Prior art methods for object recognition, motion tracking, and optical character recognition employ color probability density analysis using color histograms, such as that described in U.S. Pat. Nos. 5,845,009; 5,844,573; and 5,831,638, by Fitzpatrick, et al., Marks, et al., and Poggio, et al., respectively. Color lookup tables and display scheduling tables have been employed in prior art for displaying and interacting with graphical objects displayed with video data, such as those described in U.S. Pat. Nos. 5,835,103; 5,835,625, by Butler, et. al., and West, et al., respectively.
Many prior art methods employ traditional compression algorithms and keyword tagging methodologies to digitally characterize video streams for use with digital libraries, search and retrieval systems, and interactive television applications, such as that described in U.S. Pat. Nos. 5,832,495; 5,832,499; 5,761,417; 5,835,087; 5,845,046; 5,818,439; 5,835,667; and 5,819,286, by Gustman, Gustman, Henley, et al., Herz et al., Hirayama, et al., Nagasaka, et al., Wactlar, et al., and Yang, et al., respectively.
Other prior art methods employ cluster link generation algorithms, including node and link indices for navigating multimedia databases, such as that described in U.S. Pat. No. 5,832,494 by Egger, et al. Finally, Hyano, et al., describe a method that utilizes dot clock phase and frequency to sample color signals in video images.
All such prior art methods and systems include drawbacks which preclude wide spread acceptance for commercial applications such as video reconciliation and precise video distortion analysis. The present invention avoids the drawbacks of these methodologies and presents a novel, more commercially-viable
Blackwell Sanders Peper Martin LLP
Cooperrider F. E.
Elliott Kyle L.
Mancuso Joseph
The University of Kansas
LandOfFree
Real-time feature-based video stream validation and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Real-time feature-based video stream validation and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Real-time feature-based video stream validation and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2534191