Encoder, decoder, encoding method, and decoding method
An encoder includes circuitry and memory. Using the memory, the circuitry: encodes an original image and decodes the original image encoded, to generate a first bitstream and a local decoded image; encodes supplemental information and decodes the encoded supplemental information, to generate a second bitstream and local decoded supplemental information; inputs data based on the local decoded image and the local decoded supplemental information to a post processing network which is a neural network, to cause a reconstructed image to be output from the post processing network, the reconstructed image corresponding to the original image and being to be used to encode a following original image which follows the original image; and concatenates the first bitstream and the second bitstream to generate a concatenated bitstream.
1. An encoder, comprising:
circuitry; and
memory,
wherein, using the memory, the circuitry:
encodes an original image and decodes the original image encoded, to generate a first bitstream and a local decoded image;
encodes supplemental information and decodes the encoded supplemental information, to generate a second bitstream and local decoded supplemental information;
inputs data based on the local decoded image and the local decoded supplemental information to a post processing network which is a neural network, to cause a reconstructed image to be output from the post processing network, the reconstructed image corresponding to the original image and being to be used to encode a following original image which follows the original image; and
concatenates the first bitstream and the second bitstream to generate a concatenated bitstream.
2. The encoder according to claim 1 ,
wherein, when causing the reconstructed image to be output, the circuitry
separately inputs the local decoded image and the local decoded supplemental information as the data to the post processing network, to cause the reconstructed image to be output from the post processing network.
3. The encoder according to claim 1 ,
wherein the circuitry:
adds the local decoded image and the local decoded supplemental information; and
when causing the reconstructed image to be generated,
inputs a result of adding the local decoded image and the local decoded supplemental information as the data to the post processing network, to cause the reconstructed image to be output from the post processing network.
4. The encoder according to claim 1 ,
wherein the circuitry
subtracts the local decoded image from the original image to generate the supplemental information as a residual image.
5. The encoder according to claim 1 ,
wherein when encoding the supplemental information, the circuitry downsamples the supplemental information.
6. The encoder according to claim 5 ,
wherein when decoding the encoded supplemental information, the circuitry
upsamples the downsampled supplemental information.
7. The encoder according to claim 1 ,
wherein when encoding the supplemental information, the circuitry
inputs the supplemental information to an encoder network which is a neural network, to cause the encoded supplemental information to be output from the encoder network.
8. The encoder according to claim 7 ,
wherein, when decoding the encoded supplemental information, the circuitry
inputs the encoded supplemental information to a decoder network which is a neural network, to cause the decoded encoded supplemental information to be output from the decoder network.
9. The encoder according to claim 1 ,
wherein, when encoding the supplemental information, the circuitry
quantizes the supplemental information.
10. The encoder according to claim 1 ,
wherein, when encoding the supplemental information, the circuitry
entropy encodes the supplemental information.
11. A decoder, comprising:
circuitry; and
memory,
wherein, using the memory, the circuitry:
decodes a first bitstream included in a concatenated bitstream to generate a decoded image of codec;
decodes a second bitstream included in the concatenated bitstream to generate decoded supplemental information; and
inputs data based on the decoded image of codec and the decoded supplemental information to a post processing network which is a neural network, to cause a decoded image to be output from the post processing network.
12. The decoder according to claim 11 ,
wherein, when causing the decoded image to be output, the circuitry
separately inputs the decoded image of codec and the decoded supplemental information as the data to the post processing network, to cause the decoded image to be output from the post processing network.
13. The decoder according to claim 11 ,
wherein the circuitry:
adds the decoded image of codec and the decoded supplemental information; and
when causing the decoded image to be output,
inputs a result of adding the decoded image of codec and the decoded supplemental information as the data to the post processing network, to cause the decoded image to be output from the post processing network.
14. The decoder according to claim 11 ,
wherein, when decoding the second bitstream, the circuitry
upsamples downsampled supplemental information which is obtained from the second bitstream, to generate the decoded supplemental information.
15. The decoder according to claim 11 ,
wherein, when decoding the second bitstream, the circuitry
inputs encoded supplemental information which is obtained from the second bitstream to a decoder network which is a neural network, to cause the decoded supplemental information to be output from the decoder network.
16. The decoder according to claim 11 ,
wherein, when decoding the second bitstream, the circuitry
entropy decodes the second bitstream.
17. The decoder according to claim 11 ,
wherein, when decoding the second bitstream, the circuitry
inverse quantizes the second bitstream.
18. An encoding method, comprising:
encoding a current original image and decoding the current original image encoded, to generate a first bitstream and a local decoded image;
encoding supplemental information and decoding the encoded supplemental information, to generate a second bitstream and local decoded supplemental information;
inputting data based on the local decoded image and the local decoded supplemental information to a post processing network which is a neural network, to cause a reconstructed image to be output from the post processing network, the reconstructed image corresponding to the original image and being to be used to encode a following original image which follows the original image; and
concatenating the first bitstream and the second bitstream to generate a concatenated bitstream.
19. A decoding method, comprising:
decoding a first bitstream included in a concatenated bitstream to generate a decoded image of codec;
decoding a second bitstream included in the concatenated bitstream to generate decoded supplemental information; and
inputting data based on the decoded image of codec and the decoded supplemental information to a post processing network which is a neural network, to cause a decoded image to be output from the post processing network.