Apparatus and method of compensating image-sequences for motion

• US 5,309,237 A
• Filed: 03/31/1992
• Issued: 05/03/1994
• Est. Priority Date: 03/31/1992
• Status: Expired due to Term

• Current Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Singh, Ajit
• Primary Examiner(s): Kostak, Victor R.
• Application Number:
  US07/861,414
• Time in Patent Office: 763 Days
• Field of Search: 358/167, 358/105, 358/166, 358/37, 358/36
• US Class Current: 348/607
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: H04N 19/51 Motion estimation or motion compensation
• CPC Class Codes: H04N 19/80 Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation H04N19/635, H04N19/86 take precedence
• CPC Class Codes: H04N 5/144 Movement detection for video coding H04N19/503; analysis of motion in general G06T7/20
• CPC Class Codes: H04N 5/21 Circuitry for suppressing or minimising disturbance, e.g. moiré or halo

Abstract

The present invention apparatus and method uses a recursive filter to estimate the instantaneous intensity and temporal rate of change in intensity for object points in consecutive images of an image-sequence. The estimates of the intensity and its temporal rate of change are updated for each new image in the image sequence. The amount of smoothing introduced by the recursive filter is dependant upon the estimates computed for the intensity and temporal rate of change. As such, the recursive filter corrects the image sequence in accordance with each new set of estimates. Since motion in an image-sequence is implicit in the rate of change of intensity, the recursive filter is able to achieve motion compensated enhancements.

Claims

1. In an apparatus that utilizes an image sequence, a device for producing motion compensated enhancement for object points contained within the individual images of said image sequence, comprising:

an estimating means, including a recursive filter, for estimating an instantaneous intensity and a temporal rate of change of intensity for said object points; and

an adjusting means for adjusting said recursive filter after each image in said image sequence, according to said intensity and said temporal rate of change as estimated by said estimating means, for an immediately preceding image in said image sequence to thereby compensate said image by substantially removing noise in the presence of motion.

2. The device of claim 1 further including a dividing means for dividing said object points contained within the individual images of said image sequence into two regions, a first region wherein the intensity of said object points remain substantially unchanged between consecutive images, and a second region wherein the intensity of said object points varies between consecutive images, said estimating means, estimating said instantaneous intensity and said temporal rate of change for said object points in said first region.

3. The device of claim 2 further including a superimposing means for superimposing said first region and said second region after said instantaneous intensity and said temporal rate of change have been estimated for said object points in said first region.

4. The device of claim 3 wherein said recursive filter means includes a low-pass temporal filter.

5. The device of claim 3 wherein said recursive filter means includes a Kalman filter.

6. In an apparatus that utilizes an image sequence that contains images corrupted by noise, a device for reducing said noise in said images, comprising:

an estimating means for estimating said noise in each said image of said image sequence, wherein said estimating means estimates an instantaneous intensity and a temporal rate of change of intensity for object points contained in each said image of said image sequence; and

an adjustable recursive filter means for filtering said noise in each said image according to the value of said noise as estimated by said estimating means, for an immediately preceding image in said image sequence to thereby compensate said image by substantially removing noise in the presence of motion.

7. The device of claim 6 wherein said recursive filter includes a low-pass temporal filter.

8. The device of claim 6 wherein said recursive filter includes a Kalman filter.

9. The device of claim 8 further including a dividing means for dividing said object points, contained within each said image, into two regions, a first region wherein the intensity of said object points remain substantially unchanged between consecutive images, and a second region wherein the intensity of said object points vary between consecutive images, said recursive filter means filtering said noise form said first region.

10. The device of claim 9 further including a superimposing means for superimposing said first region and said second region after said first region has been filtered by said recursive filter means.

11. A method for providing motion-compensated enhancement for object points contained within each image of an image-sequence corrupted by noise, comprising the steps of:

separating said object points in said image-sequence into two regions, a second region wherein the intensity of said object points vary substantially between consecutive images and a first region wherein the intensity of said object points, between consecutive images, varies slowly relative to said second region;

estimating an instantaneous intensity and a temporal rate of change in intensity for said object points in each said image of said image sequence and estimating the amount of noise in said first region of each said image, using a recursive filter; and

adjusting the amount of noise reduction introduced by said recursive filter to said first region of a given image according to the amount of noise estimated for an immediately preceding image in said image sequence.

12. The method of claim 11 further including the step of superimposing said first region and said second region after said step of adjusting the amount of noise reduction introduced by said recursive filter.

13. The method of claim 12 wherein said recursive filter includes a low-pass temporal filter.

14. The method of claim 13 wherein said recursive filter includes a Kalman filter.