Adaptive filtering based upon boundary strength
View Patent ↗Adjacent regions are identified in an image. Coding parameters for the adjacent regions are identified. Selective filtering is performed at the region between the identified adjacent regions.
1. An encoder for encoding an image comprising:
an identifier for identifying a first region and a second region adjacent to each other in said image; and
a filter for selectively filtering at least a portion of said regions proximate a boundary between said regions based upon a boundary strength which is set for said boundary, wherein,
(a) a strength of said filter is controlled so that said strength increases as the boundary strength increases,
(b) said boundary strength is set to a first value when either one of said first region and said second region is intra-coded,
(c) said boundary strength is set to a second value when neither of said first region nor said second region is intra-coded and when either of said regions include non-zero coefficients,
(d) said boundary strength is set to a third value when neither of said first region nor said second region is intra-coded, when both of said regions do not include non-zero coefficients and
(1) when said regions are predicted based upon different reference images or
(2) when an absolute difference of motion vectors of said regions is equal to or greater than a non-zero threshold value in either vertical or horizontal directions,
(e) said boundary strength is set to a fourth value when neither of said first region nor said second region is intra-coded, when both of said regions do not include non-zero coefficients and
(1) when said regions are predicted based upon the same reference image and
(2) when an absolute difference of motion vectors of said regions is less than said non-zero threshold value in both vertical and horizontal directions,
(f) said first value being larger than said second value, said second value being larger than said third value and said third value being larger than said fourth value, and
(g) said fourth value indicates that there is no filtering.
2. A decoder for decoding an image comprising:
an identifier for identifying a first region and a second region adjacent to each other in said image; and
a filter, for selectively filtering at least a portion of said regions proximate a boundary between said regions based upon a boundary strength which is set for said boundary, wherein,
(a) a strength of said filter is controlled so that said strength increases as the boundary strength increases,
(b) said boundary strength is set to a first value when either one of said first region and said second region is intra-coded,
(c) said boundary strength is set to a second value when neither of said first region nor said second region is intra-coded and when either of said regions include non-zero coefficients,
(d) said boundary strength is set to a third value when neither of said first region nor said second region is intra-coded, when both of said regions do not include non-zero coefficients and
(1) when said regions are predicted based upon different reference images or
(2) when an absolute difference of motion vectors of said regions is equal to or greater than a non-zero threshold value in either vertical or horizontal directions,
(e) said boundary strength is set to a fourth value when neither of said first region nor said second region is intra-coded, when both of said regions do not include non-zero coefficients and
(1) when said regions are predicted based upon the same reference image and
(2) when an absolute difference of motion vectors of said regions is less than said non-zero threshold value in both vertical and horizontal directions,
(f) said first value being larger than said second value and said second value being larger than said third value and said third value being larger than said fourth value, and
(g) said fourth value indicates that there is no filtering.