Video image processing
AssigneeBritish Broadcasting Corporation
InventorMichael Burl
Life of this patent
Estimating the term…
Abstract
In a digital video display system, a composite time-compensated video picture, generated from a plurality of input fields, includes discontinuities, e.g. aliases resulting from relative movement of a foreground object and a background. These discontinuities are the results of time-projections from the sequential input fields or frames, which include movement of objects between the respective instants of the frames. These discontinuities are detected and are smoothed by a softening filter, in order to improve the quality of the resulting output image. In a preferred embodiment, the filtering is applied directionally, based on the nature of the discontinuity detected.
Claims11
A method of improving a video image by detecting, and compensating for, discontinuities in a motion-compensated composite picture generated from a plurality of input fields, wherein the discontinuities arise from differential motions between adjacent areas of an input field, comprising the steps of:
- assigning motion vectors to elementary areas of each of a plurality of input fields in a sequence of video images;
- generating a shift field from a time projection of the motion vectors assigned to at least one of the plurality of input fields in the sequence of video images,
- detecting transitions in displacement in the shift field between adjacent areas of the shift field, and
- disguising, by use of a filter, any transition detected.
A method according to any one of the preceding claims, further comprising the step of detecting the direction of any discontinuity which is detected.
An apparatus for improving a video image by detecting, and compensating for, discontinuities in a motion-compensated composite picture generated from a plurality of input fields wherein the discontinuities arise from differential motion between adjacent areas of an input filed comprising
- means for assigning motion vectors to elementary areas of each input field in a sequence of video images;
- means for generating a shift field from a time projection of the motion vectors assigned to at least one of the plurality of input fields,
- means for detecting transitions in the vectors of the shift field between adjacent areas of the shift field and
- means for disguising any transition detected by filtering at that transition, prior to display of a motion-compensated picture.
Dependencies are inferred from the claim text.