Encoder, decoder, encoding method, and decoding method
An encoder includes circuitry and memory. The circuitry, using the memory: derives a one-dimensional array of a plurality of reference samples for intra prediction; performs smoothing on the one-dimensional array of the plurality of reference samples which has been derived; and generates a prediction image using the plurality of reference samples. In deriving the one-dimensional array, the circuitry projects a value of at least one decoded pixel located on a first line onto a second line perpendicular to the first line, to derive at least one of the plurality of reference samples, and the smoothing is performed on the at least one decoded pixel projected onto the second line.
1. An encoder which encodes a video, comprising:
circuitry; and
memory, wherein
using the memory, the circuitry:
derives a one-dimensional array of a plurality of reference samples for intra prediction by projecting a value of at least one first decoded pixel located on a first line onto a second line to derive the one-dimensional array of the plurality of reference samples, the second line being perpendicular to the first line and including values of a plurality of second decoded pixels, and the plurality of reference samples of the one- dimensional array being located on the second line and including the value of the at least one first decoded pixel projected on the second line and the values of the plurality of second decoded pixels;
after deriving the one-dimensional array of the plurality of reference samples, performs smoothing on the derived one-dimensional array of the plurality of reference samples; and
after performing smoothing on the derived one-dimensional array of the plurality of reference samples, generates a prediction image using the plurality of reference samples upon which the smoothing has been performed,
wherein the circuity performs the smoothing on the derived one-dimensional array of the plurality of reference samples according to the following:
smoothed
-
ref
[
x
]
=
{
(
ref
[
x
-
1
]
+
2
*
ref
[
x
]
+
ref
[
x
+
1
]
)
⪢
2
,
(
min_
x
+
1
)
<
x
<
(
max_
x
-
1
)
(
3
*
ref
[
x
]
+
ref
[
x
+
1
]
)
⪢
2
,
x
=
min_
x
(
ref
[
x
-
1
]
+
3
*
ref
[
x
]
)
⪢
2
,
x
=
max_
x
,
where ref[x] is a reference sample at index x from among the derived one-dimensional array of the plurality of reference samples, min_x is the minimum value of the indexes of the derived one-dimensional array of the plurality of reference samples, max_x is the maximum value of the indexes of the derived one-dimensional array of the plurality of reference samples, smoothed_ref[x] is a smoothed reference sample at index x from among the plurality of reference samples upon which the smoothing has been performed.
2. The encoder according to claim 1 , wherein
the circuitry performs the smoothing on the derived one-dimensional array of the plurality of reference samples immediately before calculating a prediction signal based on the plurality of reference samples upon which the smoothing has been performed.
3. The encoder according to claim 1 , wherein
the smoothing is performed by applying a one-dimensional spatial smoothing filter to the derived one-dimensional array of the plurality of reference samples.
4. The encoder according to claim 3 , wherein
the one-dimensional spatial smoothing filter is applied at a boundary of the derived one-dimensional array of the plurality of reference samples, by extending the derived one-dimensional array using a value of at least one reference sample located at the boundary among the plurality of reference samples.
5. The encoder according to claim 1 , wherein
in generating the prediction image, the circuitry performs interpolation processing by applying an interpolation filter to the derived one-dimensional array of the plurality of reference samples upon which the smoothing has been performed, and
the interpolation filter is applied regardless of whether a position of any of the plurality of reference samples of the derived one-dimensional array is at an integer position or at a non-integer position.
6. A decoder, comprising:
circuitry; and
memory, wherein
using the memory, the circuitry:
derives a one-dimensional array of a plurality of reference samples for intra prediction by projecting a value of at least one first decoded pixel located on a first line onto a second line to derive the one-dimensional array of the plurality of reference samples, the second line being perpendicular to the first line and including values of a plurality of second decoded pixels, and the plurality of reference samples of the one-dimensional array being located on the second line and including the value of the at least one first decoded pixel projected on the second line and the values of the plurality of second decoded pixels;
after deriving the one-dimensional array of the plurality of reference samples, performs smoothing on the derived one-dimensional array of the plurality of reference samples; and
after performing smoothing on the derived one-dimensional array of the plurality of reference samples, generates a prediction image using the plurality of reference samples upon which the smoothing has been performed,
wherein the circuity performs the smoothing on the derived one-dimensional array of the plurality of reference samples according to the following:
smoothed
-
ref
[
x
]
=
{
(
ref
[
x
-
1
]
+
2
*
ref
[
x
]
+
ref
[
x
+
1
]
)
⪢
2
,
(
min_
x
+
1
)
<
x
<
(
max_
x
-
1
)
(
3
*
ref
[
x
]
+
ref
[
x
+
1
]
)
⪢
2
,
x
=
min_
x
(
ref
[
x
-
1
]
+
3
*
ref
[
x
]
)
⪢
2
,
x
=
max_
x
,
where ref[x] is a reference sample at index x from among the derived one-dimensional array of the plurality of reference samples, min_x is the minimum value of the indexes of the derived one-dimensional array of the plurality of reference samples, max_x is the maximum value of the indexes of the derived one-dimensional array of the plurality of reference samples, smoothed_ref[x] is a smoothed reference sample at index x from among the plurality of reference samples upon which the smoothing has been performed.
7. The decoder according to claim 6 , wherein
the circuitry performs the smoothing on the derived one-dimensional array of the plurality of reference samples immediately before calculating a prediction signal based on the plurality of reference samples upon which the smoothing has been performed.
8. The decoder according to claim 6 , wherein
the smoothing is performed by applying a one-dimensional spatial smoothing filter to the derived one-dimensional array of the plurality of reference samples.
9. The decoder according to claim 8 , wherein
the one-dimensional spatial smoothing filter is applied at a boundary of the derived one-dimensional array of the plurality of reference samples, by extending the derived one-dimensional array using a value of at least one reference sample located at the boundary among the plurality of reference samples.
10. The decoder according to claim 6 , wherein
in generating the prediction image, the circuitry performs interpolation processing by applying an interpolation filter to the derived one-dimensional array of the plurality of reference samples upon which the smoothing has been performed, and
the interpolation filter is applied regardless of whether a position of any of the plurality of reference samples of the derived one-dimensional array is at an integer position or at a non-integer position.