Image analysis – Image enhancement or restoration – Image filter
Reexamination Certificate
2001-02-02
2004-07-20
Werner, Brian (Department: 2621)
Image analysis
Image enhancement or restoration
Image filter
C382S275000, C382S300000
Reexamination Certificate
active
06766063
ABSTRACT:
BACKGROUND
Nonlinear editing systems commonly process video that is stored in a subsampled component video format. There are a number of subsampled component video formats, such as YCrCb 4:2:2 video data. Data in this format may be received as RGB component video data that is converted into YCrCb data in a 4:4:4 format, which is then decimated to the subsampled 4:2:2 format.
Graphics data that is imported into a nonlinear editing system commonly is in an RGB component video data format. To avoid various visible artifacts in an image, due particularly to aliasing, graphics data commonly is filtered prior to decimation.
Graphics data also is commonly exported from a nonlinear editing system from the subsampled YCrCb 4:2:2 format into the RGB component video data format. This export process involves interpolating or filtering the decimated data to produce the RGB component video data.
SUMMARY
If graphics data is imported into and then exported from a nonlinear editing system several times, image quality tends to degrade due to multiple filtering operations applied to the graphics data on import. Sometimes, a portion of graphics data imported into a nonlinear editing system has been processed by a nonlinear editing system, whereas another portion has not. To avoid such degradation of image quality, portions of graphics data that have been previously filtered and decimated are identified and not filtered prior to decimation. Other portions of graphics data are filtered and decimated.
Accordingly, in one aspect, an image in a first color representation is imported into digital nonlinear video editing system that uses images in a second subsampled color representation by detecting in a portion of the image in the first color representation a signature of an upsampling filter for generating an image in the first color representation from an image in the second subsampled color representation. If the signature is detected, the portion of the image in the first color representation is decimated without filtering the portion of the image. If the signature is not detected, the portion of the image in the first color representation is filtered before decimating the portion of the image.
In another aspect, an image in a first color representation is imported into digital nonlinear video editing system that uses images in a second subsampled color representation by determining for each pixel in an image in the second color representation, whether a corresponding portion of the image in the first color representation was filtered using an upsampling filter for generating an image in the first color representation from an image in the second subsampled color representation. If the portion of the image in the first color representation was filtered using the upsampling filter, the portion of the image is decimated without filtering the portion of the image. If the portion of the image in the first color representation was not filtered using the upsampling filter, the portion of the image is filtered before decimating the portion of the image.
Another aspect is a method or system for determining a measure of likelihood that an image in a first color representation was filtered using an upsampling filter used by a digital nonlinear video editing system to generate an image in the first color representation from an image in a second subsampled color representation. A set of pixels in the image in the first color representation is selected according to the second subsampled color representation. A value for each of a plurality of pixels intermediate the selected pixels is estimated. The estimated value is compared to an actual value for the plurality of pixels intermediate the selected pixels in the image in the first color representation.
Another aspect is a method or system for generation adaptive filtering an image in a first color representation during conversion of the image to a second subsampled color representation used in a digital nonlinear video editing system. Information is received that is indicative of, for each of a plurality of selected pixels in the image in the first color representation, whether the portion of the image around the selected pixel was filtered using an upsampling filter used by the digital nonlinear video editing system to generate an image in the first color representation from an image in the second subsampled color representation. Each of the selected pixels of the image in the first color representation is selectively filtered and decimated according to whether the portion of the image around the pixel was filtered using the upsampling filter.
Another aspect is a digital information product that includes a computer readable medium and information stored on the computer readable medium that, when interpreted by a computer, indicates, for each of a plurality of selected pixels in an image in a first color representation, whether that the portion of the image around the selected pixel was filtered using an upsampling filter used by a digital nonlinear video editing system to generate an image in the first color representation from an image in a second subsampled color representation.
In the various foregoing aspects, detection may be performed by determining a measure of likelihood that an image in a first color representation was filtered using an upsampling filter used by a digital nonlinear video editing system to generate an image in the first color representation from an image in a second subsampled color representation. The measure of likelihood may be determined by selecting a set of pixels in the image in the first color representation according to the second subsampled color representation, estimating a value for each of a plurality of pixels intermediate the selected pixels, and comparing the estimated value to an actual value for the plurality of pixels intermediate the selected pixels in the image in the first color representation. Such comparison may be performed by determining whether the estimated value is within a range defined around the actual value. In the various foregoing aspects, filtering may be performed by applying a tent filter to pixel data comprising the portion of the image.
REFERENCES:
patent: 6453074 (2002-09-01), Zheng
patent: 6496186 (2002-12-01), Deering
Moghaddamzadeh, Ali et al., “An Object-Based Approach to Color Subsampling”, 1996 IEEE, No. 0-7803-3192-3/96, pp. 1894-1897.
Murching, Anil M., et al., “Adaptive Color Subsampling Of Images”, Proc.SPIE Conf. Visual Communications and Image Proc., vol. 2094, Part 3, pp. 1198-1206, Nov. 1993.
Murching, Anil M., et al., “Adaptive Subsampling of Color Images”, Proceedings ICIP 1994, vol. 3.
Turkowski, Ken, “Filters for Common Resampling Tasks”, Apr. 10, 1990.
Avid Technology Inc.
Edwards Patrick
Gordon Peter J.
Werner Brian
LandOfFree
Generation adaptive filtering for subsampling component... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Generation adaptive filtering for subsampling component..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Generation adaptive filtering for subsampling component... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3224596