Video signal noise reduction system using time-varying filter coefficients

• US 5,442,407 A
• Filed: 03/22/1994
• Issued: 08/15/1995
• Est. Priority Date: 03/22/1994
• Status: Expired due to Term

• Current Assignee: Panasonic Corporation Of North America (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Corporation Of America (Panasonic Holdings Corporation)
• Inventors: Iu, Siu-Leong
• Primary Examiner(s): Groody, James J.
• Assistant Examiner(s): Murrell, Jeffrey S.
• Application Number:
  US08/215,683
• Time in Patent Office: 511 Days
• Field of Search: 348/616, 348/617, 348/619, 348/618, 348/620, 348/621, 348/415, 348/416, 348/420, 348/421
• US Class Current: 348/620
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: H04N 5/21 Circuitry for suppressing or minimising disturbance, e.g. moiré or halo

Abstract

A video signal noise reduction system is disclosed in which corresponding image pixels from a current frame and image pixels from a noise reduced previous frame are averaged to produce respective noise reduced pixel values. In the disclosed exemplary embodiment of the invention, a processed frame memory stores a noise reduced image and a motion estimator identifies a target block of pixels in this processed frame memory to be used in calculating a noise reduced video signal. A motion compensated noise reduction processor combines the block of the noise reduced image pixels stored in the processed frame memory with a current block of pixels generated from the video signal to produce a motion block of noise reduced pixels. The relative weighting applied to the stored pixel and the current pixel depends on the number of pixels that were used to generate the stored pixel. A breakdown detector discards pixels in the motion block if a measure of noise reduction achieved for the one component divided by a measure of noise in the motion block is greater than a threshold value. Sparkle reduction is used to replace any original component of the motion block with its corresponding noise reduced component if all of the components adjacent to the original component have been replaced by their corresponding noise reduced components.

Claims

1. Apparatus for reducing noise in an input video signal which includes a plurality of video frames, the apparatus comprising:

processed frame memory means for storing a noise reduced image;

motion estimator means for selecting a target block of image components in said frame memory to be used in calculating a noise reduced video signal; and

motion compensated noise reduction means for calculating said noise reduced video signal by calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced image components by averaging a plurality of image components in a current block selected from the video signal with respective image components of the target block as weighted by the coefficients.

2. Apparatus for reducing noise in an input video signal which includes a plurality of video frames, the apparatus comprising:

processed frame memory means for storing a noise reduced image;

motion estimator means for identifying a target block of image components in said frame memory to be used in calculating a noise reduced video signal;

motion compensated noise reduction means for calculating said noise reduced video signal by calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced image components by averaging a plurality of image components in a current block selected from the video signal with respective image components of the target block as weighted by the coefficients; and

breakdown detection means for replacing each one image component of the current block with a corresponding noise-reduced image component of the motion block if a measure of noise reduction achieved for the one image component divided by a reference measure of noise is less than a threshold value.

3. Apparatus according to claim 2, further comprising sparkle reduction means for examining the current block to identify image components which have not been replaced by noise reduced image components and replacing the identified image components with corresponding noise reduced image components from the motion block if all of the image components which are adjacent to the identified image components have been replaced.

4. Apparatus according to claim 2, further comprising:

count memory means such that a 1;

1 correspondence exists between counts and image components in one frame of the plurality of frames, where each count indicates a number of image components corresponding to the count in the plurality of frames which were used to develop the noise reduced image component;

wherein the motion compensated noise reduction means is responsive to the counts held in the count memory means and corresponding to the respective image components of the target block for calculating the set of time-varying weighted filtering coefficients for the target block.

5. Apparatus for reducing noise in an input video signal which includes a plurality of video frames, the apparatus comprising:

processed frame memory means for storing a noise reduced image;

motion estimator means for identifying a target block of image components in said frame memory to be used in calculating a noise reduced video signal;

motion compensated noise reduction means for calculating said noise reduced video signal by calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced image components by averaging a plurality of image components in a current block selected from the video signal with respective image components of the target block as weighted by the coefficients; and

classification means for determining a measure of variation which occurs in the current block and for calculating a weighting value and a threshold value to be used by the motion estimator means.

6. Apparatus for reducing noise in an input video signal which includes a plurality of video frames, the apparatus comprising:

processed frame memory means for storing a noise reduced image;

motion estimator means for identifying a target block of image components in said frame memory to be used in calculating a noise reduced video signal;

motion compensated noise reduction means for calculating said noise reduced video signal by calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced image components by averaging a plurality of image components in a current block selected from the video signal with respective image components of the target block as weighted by the coefficients; and

means for providing a plurality of trial velocity vectors to the motion estimator means for use in identifying the target block.

7. Apparatus for reducing noise in an input video signal which includes a plurality of video frames, the apparatus comprising:

processed frame memory means for storing a noise reduced image;

motion estimator means for identifying a target block of image components in said frame memory to be used in calculating a noise reduced video signal;

motion compensated noise reduction means for calculating said noise reduced video signal by calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced image components by averaging a plurality of image components in a current block selected from the video signal with respective image components of the target block as weighted by the coefficients; and

future frame memory means for holding image components representing an image which is to be displayed following the current image, wherein the motion estimator means includes means for identifying a further block of image components in said future frame memory means and the motion compensated noise reduction means includes means for generating said motion block of image components by combining the current block of image components with at least one of the target block of image components and the further target block of image components.

8. A method of reducing noise in a video signal which includes a plurality of video frames, the method comprising the steps of:

storing a noise reduced image in processed frame memory;

selecting a target block of picture elements in said processed frame memory to be used in calculating a noise reduced video signal; and

calculating said noise reduced video signal by combining the block of the noise reduced image pixels stored in the processed frame memory with a current block of pixels generated from the video signal and calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced pixels by averaging the plurality of pixels in the block of noise reduced image pixels with respective pixels of the current block of pixels as weighted by the coefficients.

9. A method of reducing noise in a video signal which includes a plurality of video frames, the method comprising the steps of:

storing a noise reduced image in a processed frame memory;

identifying a target block of picture elements in said processed frame memory to be used in calculating a noise reduced video signal; and

calculating said noise reduced video signal by combining the block of the noise reduced image pixels stored in the processed frame memory with a current block of pixels generated from the video signal and calculating a set of time-varying weighted filtering coefficients for the target block and for generating a motion block of noise reduced pixels by averaging the plurality of pixels in the block of noise reduced image pixels of the current block of pixels as weighted by the coefficients; and

replacing each one of the pixels of the current block with a corresponding noise-reduced pixel from the motion block if a measure of noise reduction achieved for the one pixel divided by a measure of noise in the video signal is less than a threshold value.

10. A method according to claim 9 further comprising the step of examining the current block to identify pixels which have not been replaced by noise-reduced pixels and replacing the identified pixels with corresponding noise reduced pixels of the motion block if all of the image pixels which are adjacent to the identified pixels have been replaced.

11. A method according to claim 9 further comprising the step of determining a measure of the variation which occurs in the current block of pixels and for calculating a weighting value and a threshold value to be used in motion estimation and breakdown detection.

12. A method according to claim 11, wherein the step of determining a measure of variation occurring in the current block of pixels includes the step of identifying and counting image pixels which represent edges in the current block of pixels.

13. A method according to claim 9 further comprising the step of providing a plurality of trial velocity vectors to the motion estimator means for use in identifying the target block of pixels.