Noise adaptive 3D composite noise reduction
A method for filtering digital video is disclosed. The method generally includes the steps of (A) checking a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of the artifacts is based on an adjustable threshold, (B) first filtering each of the original blocks having at least one of the artifacts to remove the at least one artifact and (C) second filtering each of the original blocks lacking at least one of the artifacts to remove noise, wherein the second filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of the original blocks.
1. A method for filtering digital video, comprising the steps of:
(A) checking a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold;
(B) first filtering each of said original blocks having at least one of said artifacts to remove said at least one artifact;
(C) second filtering each of said original blocks lacking at least one of said artifacts to remove noise, wherein said second filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks;
(D) measuring said respective noise levels for said original blocks;
(E) detecting each saturated block of said original blocks in which most of a plurality of pixel values comprise saturated values; and
(F) excluding said saturated blocks from said measuring of said respective noise levels.
2. The method according to claim 1 , further comprising the step of:
measuring said respective noise levels for said original blocks based on a plurality of sum-of-absolute-differences values between said original blocks and a plurality of previous blocks in a previous field, wherein said previous field has a same parity as said original field.
3. The method according to claim 1 , further comprising the step of:
measuring said respective noise levels for said original blocks based on a plurality of intra-block activities of said original blocks.
4. The method according to claim 1 , further comprising the step of:
measuring said respective noise levels for said original blocks based on a plurality of sum-of-absolute-differences values between a plurality of even line in said original blocks and a plurality of odd lines in a plurality of previous blocks in a previous field, wherein said previous field has a same parity as said original field.
5. The method according to claim 1 , further comprising the steps of:
measuring an activity in each of said original blocks; and
generating said adjustable threshold based on said activity.
6. The method according to claim 1 , further comprising the steps of:
measuring a random noise in said original field; and
generating said adjustable threshold based on said random noise.
7. The method according to claim 1 , further comprising the steps of:
measuring a strength of dot crawl for each of said original blocks; and
adjusting said second filtering based on said strength.
8. The method according to claim 1 , further comprising the steps of:
detecting each stationary block of said original blocks; and
performing said checking of said original blocks for said artifacts only on said stationary blocks.
9. The method according to claim 1 , further comprising the step of:
temporal smoothing a noise estimate used in said second filtering, wherein
said first filtering is performed for both (i) a plurality of luminance blocks of said original field and (ii) a plurality of chrominance blocks of said original field, and
said second filtering is performed on all of said original blocks after said first filtering.
10. A system comprising:
a first processor configured to check a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold; and
a second processor configured to (i) artifact filter each of said original blocks having at least one of said artifacts to remove said at least one artifact and (ii) noise filter each of said original blocks lacking at least one of said artifacts to remove noise, wherein (A) said noise filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks and (B) said second processor is further configured to (i) measure said respective noise levels for said original blocks, (ii) detect each saturated block of said original blocks in which most of a plurality of pixel values comprise saturated values and (iii) exclude said saturated blocks from said measuring of said respective noise levels.
11. The system according to claim 10 , wherein said second processor is further configured to measure said respective noise levels for said original blocks based on temporal noise differences between said original blocks and a plurality of previous blocks in a previous field, wherein said previous field has a same parity as said original field.
12. The system according to claim 10 , wherein said second processor is further configured to measure said respective noise levels for said original blocks based on spatial intra-block activities of said original blocks.
13. The system according to claim 10 , wherein said second processor is further configured to measure said respective noise levels for said original blocks based on cross spatio-temporal differences between a plurality of even line in said original blocks and a plurality of odd lines in a plurality of previous blocks in a previous field, wherein said previous field has a same parity as said original field.
14. The system according to claim 10 , wherein said second processor is further configured to (i) measure an activity in each of said original blocks and (ii) generate said adjustable threshold based on said activity.
15. The system according to claim 10 , wherein said second processor is further configured to (i) measure a random noise in said original field and (ii) generate said adjustable threshold based on said random noise.
16. The system according to claim 10 , wherein said second processor is further configured to (i) measure a strength of dot crawl for each of said original blocks and (ii) adjust said noise filtering based on said strength.
17. The system according to claim 10 , wherein said second processor is further configured to (i) detect each stationary block of said original blocks and (ii) perform said checking of said original blocks for said artifacts only on said stationary blocks.
18. A system comprising:
means for checking a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold;
means for first filtering each of said original blocks having at least one of said artifacts to remove said at least one artifact;
means for second filtering each of said original blocks lacking at least one of said artifacts to remove noise, wherein said second filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks; means for measuring said respective noise levels for said original blocks;
means for detecting each saturated block of said original blocks in which most of a plurality of pixel values comprise saturated values; and
means for excluding said saturated blocks from said measuring of said respective noise levels.
19. A method for filtering digital video, comprising the steps of:
(A) checking a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold;
(B) first filtering each of said original blocks having at least one of said artifacts to remove said at least one artifact;
(C) second filtering each of said original blocks lacking at least one of said artifacts to remove noise, wherein said second filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks;
(D) detecting each stationary block of said original blocks; and
(E) performing said checking of said original blocks for said artifacts only on said stationary blocks.
20. A method for filtering digital video, comprising the steps of:
(A) checking a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold;
(B) first filtering each of said original blocks having at least one of said artifacts to remove said at least one artifact;
(C) second filtering each of said original blocks lacking at least one of said artifacts to remove noise, wherein said second filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks; and
(D) temporal smoothing a noise estimate used in said second filtering, wherein (i) said first filtering is performed for both (a) a plurality of luminance blocks of said original field and (b) a plurality of chrominance blocks of said original field, and (ii) said second filtering is performed on all of said original blocks after said first filtering.
21. A system comprising:
a first processor configured to check a plurality of original blocks in an original field for a plurality of artifacts, wherein detection of said artifacts is based on an adjustable threshold; and
a second processor configured to (i) artifact filter each of said original blocks having at least one of said artifacts to remove said at least one artifact, (ii) noise filter each of said original blocks lacking at least one of said artifacts to remove noise, (iii) detect each stationary block of said original blocks and (iv) perform said checking of said original blocks for said artifacts only on said stationary blocks, wherein said noise filtering is (i) motion compensated and (ii) adaptive to a respective noise level in each of said original blocks.