Video image processing

Image analysis – Image enhancement or restoration – Artifact removal or suppression

Patent

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

348452, G06K 940

Patent

active

056339569

DESCRIPTION:

BRIEF SUMMARY
FIELD OF THE INVENTION

This invention relates to video image processing and in particular to picture creation with motion compensation image processing.


BACKGROUND

Various aspects of motion compensation are described in our British Patent Applications Nos. 9020498.3 and 9111348.0 which describe systems capable of generating intermediate fields of video image sequences where the original sequence of images includes motion of, for example, a foreground object moving over a background. Such intermediate fields are required when producing slow motion sequences or for standards conversion.
In many applications of motion compensation an output image is created by cutting and pasting one or more parts of various video images onto a background. The parts of the video images may include motion within themselves and the parts as a whole may be moving over the background.
When such cutting and pasting techniques are used visible discontinuities are often produced in the output image when there is a sharp transition from foreground image to background image. These discontinuities are referred to as `aliases` and the problem of these is well known in picture creation systems. They are removed from pictures by use of `anti-aliasing` techniques which blend the foreground and background pictures together over a few pixels.


SUMMARY OF THE INVENTION

Preferred embodiments of the present invention provide a system which detects these discontinuities in ouput images and then blends the foreground and background images together over a few pixels.


BRIEF FIGURE DESCRIPTION

The invention will now be described in detail by way of example with reference to the accompanying drawings in which:
FIG. 1 shows two input fields from an original video sequence with an Intermediate ouput field; and
FIG. 2 shows circuit for removing discontinuities from the output field.
FIG. 3 shows a more complex circuit embodying the invention.


DETAILED DESCRIPTION

A discontinuity does not occur if an area of the picture is simply shifted, i.e. all of the image was moved by the same amount.
A discontinuity will however occur if the shift between two adjacent elements of a picture is different.
The shift applied to a picture element (pixel) is the motion velocity of that pixel multiplied by time between the input and output pictures. In regions of obscured and revealed background i.e. moving foreground object, the time vector is distorted so that parts of the image are taken from either the input field before or the input field after the output field time. This is called the cross-fade signal.
In a two field motion compensation algorithm the output picture is made from two sequential input fields and two motion vector fields. The two vector fields are associated with the forward projection of the earlier input fields and the backward projection of the following picture field.
The first step in identifying a discontinuity in an output field is to calculate the displacement from the velocity vectors multiplied by the distance in time. This displacement is termed the shift field and is then passed to a small two dimensional differentiating filter to look for transitions in the shift field. If the output of the filter is above a predetermined threshold then a discontinuity has been found. The threshold is set at a level of say half a pixel. A low pass filter is then applied to the discontinuity to soften the transition between the two pictures. This softening is equivalent to a blending of the two pictures by taking different proportions of each image when generating a pixel.
The softening filter only operates on significant fractures in the picture. Tiny differences between various measurements of the same motion are invisible in the output picture and the unnecessary use of a blurring filter may do more damage than good.
FIG. 1 shows the situation of a foreground object moving upwards and the background moving downwards. Several different situations arise and these are described below: contribute to the output picture. picture is taken from F1 on

REFERENCES:
patent: 4651207 (1987-03-01), Bergmann et al.
patent: 4864394 (1989-09-01), Gillard
patent: 5068727 (1991-11-01), Haghiri et al.
patent: 5134480 (1992-07-01), Wang et al.
"A Combination of DPCM and Motion-Compensated Frame Interpolation for the Encoding of 34 Mbit/s Colour TV Signals", Frequenz, vol. 42, No. 8, pp. 217-222, Grotz et al., Aug. 1988.
"Coding Color Image Sequences at 64 kbit/s--A New Approach for an ISDN-Videophone (Part III)" Frequenz, vol. 43, No. 5, pp. 126-133, Gilge et al., May, 1989.

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Video image processing does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Video image processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video image processing will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-2334932

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.