Television – Image signal processing circuitry specific to television – Noise or undesired signal reduction
Patent
1997-10-24
2000-05-30
Britton, Howard
Television
Image signal processing circuitry specific to television
Noise or undesired signal reduction
348416, 348443, 348699, 708313, H04N 726
Patent
active
060696702
DESCRIPTION:
BRIEF SUMMARY
This invention relates to improved motion compensated filtering or interpolation of interlaced signals and in particular video sequences such as television pictures, with application in all areas where interlaced video sequences are processed using motion vectors, including but not limited to:
In this application the term video sequence refers to an image sequence sampled in an interlaced fashion. An image may be a multi-dimensional representation of some property-typically, but not necessarily brightness-of a scene; thus a moving image is a temporal and spatial representation of a property of a scene and an image sequence is a moving image sampled at specific instants. The instant at which a point in the moving image is sampled may depend on the spatial location of that point. A pixel is a sample of an image having a value corresponding to some property of the scene which the image represents and corresponding to a specific image co-ordinate. The image co-ordinate, a specification of the location of a point in an image, need not correspond to a point at which a sampled image has been sampled. Hence image co-ordinates, input and output, are not necessarily quantised. A sampling site is the image co-ordinate of a pixel in a sampled image.
Conventional motion compensated filtering or interpolation is described in detail in references 2, 3, 4, 5, 10, 12, 13, 14, 16, 17, 18 & 20 listed in appendix 1. This process will henceforth be referred to as MCF, or conventional MCF, which will be assumed to include the process of interpolation as well as filtering.
The hardware implementation of a conventional motion compensated filter/interpolator is described in references 3, 4, 18 & 20. One simple implementation is as a cascade of filter/interpolation cells or taps as described in reference 3 and illustrated in FIG. 1.
FIG. 1 shows a conventional MCF filter cell having a vector processor 1 which converts the motion vectors for each output sampling site in an output image into the input coordinates of an input image used in the filtering or interpolation operation.
The vector processor receives the output sampling site, generated by counters as described in the references, and the associated motion vector from motion estimator hardware, described in the references, and searches through consecutive fields to find the correct input co-ordinate corresponding to the output sampling site concerned.
The motion estimator is a device for estimating the motion vectors corresponding to a moving image. The estimated motion vectors can be sampled on either the input or output sampling lattices for an image processing system. For MCF input and output sampling lattices are usually, identical. For motion compensated interpolation, the estimated motion vectors should be sampled on the output sampling lattice.
The input coordinate depends on the output sampling site, the motion vector, and the location of the filter tap within the filter aperture. A variable delay 2 supplies a multiplier 3 with the correct input pixel value, from an input image to an image processing system, for the current output sampling site which is determined by the integer part of the current input coordinate. The multiplier 3 multiplies the value of the pixel by a pre-calculated coefficient held in a look up table 4 stored in ROM. The coefficient is determined by the required filter characteristics and the sub-pixel, or fractional, part of the input coordinate supplied by the vector processor as shown in FIG. 1.
The term filture aperture in this document refers to the impulse response of the filter. The apertue is distinct from the filter's region of support which describes the number and spatio-temporal distribution of the filter taps. A typical motion compensated transversal filter/interpolator, such as that used for television standards conversion, might have an aperture spanning four fields, 4 field lines on each field and 2 pixels on each field line. Field line spacing is twice the picture line spacing which is the distance between vertically consecutive li
REFERENCES:
patent: 4862266 (1989-08-01), Gillard
patent: 4862267 (1989-08-01), Gillard
patent: 5335194 (1994-08-01), Clayton
"Signal Processing of HDTV", edited by L. Chiariglione. Proceeding of the Second International Workshop on Signal Processing of HDVT. L'Aquila, Italy, Feb. 29-Mar. 2, 1998.
"Motion Compensated Subsampling of HDTV", by R.A.F. Belfor, R.L. Lagendijk, and J. Biemond. SPIE Proceedings, vol. 1605. Visual Communications and Image Processing '91: Visual Communication, pp. 274-284, Nov. 11, 1991.
"Motion Compensated Field Interpolation Using A Hierarchically Structured Displacement Estimator", by Matthias Bierling and Robert Thoma. Signal Processing, vol. 11, No. 4, Dec., 1986, pp. 387-404.
"Television Standards Conversion", by Timothy Borer. A thesis submitted for the degree of Doctor of Philosophy. Supervised by Josef Kittler, Department of Electronic and Electrical Engineering, University of Surrey, Guilford, Surrey GU2 5XH, U.K., and Ian Childs, Research Dept., British Broadcasting Corporation, Kingswood Warren, Tadworth, Surrey, KT20 6NP, U.K.
"Motion Compensated Display Field Rate Upconversion", by T.J. Borer, M.G. Huyler & D.W. Parker. International Broadcasting Convention, Birghten, UK. Sep. 21-25, 1990.
"Pre and Post-Processing in a Video Terminal Using Motion Vectors" by L. Chiariglione, L. Corgnier, M. Guglielmo. Centro Studi e Laboratori Telecomunicazioni, Italy.
"Digital Standards Conversion: interpolation theory and aperture systhesis" by C.K.P. Clarke, A.C.G.I. & N.E. Tanton.Research Department, Engineering Division, British Broadcasting Corporation, Dec. 1984.
"Review of Techniques for Motion Estimation and Motion Compensation" by Eric Dubois & Janusz Konrad. 4th Int'l Colloquium on Advanced Television Systems, Ottawa, Canada, Jun. 25-29, 1990.
"Motion-compensating Field Interpolation from Interlaced and Non-interlaced Grids" by B.Girod & R. Thoma 2nd Int'l Technical Symposium on Optical & Electro-Optical Applied Science & Engineering: Image Processing Symposium, Cannes, Dec., 1985.
"Motion Compensated Interpolation" by Elena Marcozzi & Stefano Tubaro. SPIE vol. 804, Advances in Image Processing (1987). pp. 61-68.
"Standards Conversion Using Motion Compensation" by Thomas Reuter. Signal Processing, vol. 16 (1989) pp. 73-82.
"Generation of High Quality Slow-Motion Replay Using Motion Compensation" by G.A. Thomas & H.Y.K. Lau. International Broadcasting Convention, Sep. 21-25, 1990. British Broadcasting Corp. U.K.
"Motion Compensated Interpolation Considering Covered and Uncovered Background" by Robert Thoma & Matthias Bierling. Signal Processing: Image Communication, vol. 1, No. 2, Oct., 1989.
"Television Motion Measurement for DATV and Other Applications" by G.A. Thomas. Research Department Report, Sep., 1987. The British Broadcasting Corp.
"Distorting the Time Axis: Motion Compensated Image Processing in the Studio" by G.A. Thomas. BBC Research Dept. International Broadcasting Convention, Brighton, UK, Sep. 23-27, 1988.
"Review of Motion Analysis Techniques" by J.F. Vega-Riveros & K. Jabbour. IEE Proceedings, vol. 136, Pt. 1, No. 6, Dec., 1989.pp. 397-404.
Britton Howard
Innovision Limited
LandOfFree
Motion compensated filtering does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Motion compensated filtering, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Motion compensated filtering will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1914331