Image decoding method and apparatus using same

Patent metadata

Field	Value
Publication number	US-10165282-B2
Application number	US-201314381454-A
Country	US
Kind code	B2
Filing date	Nov 21, 2013
Priority date	Nov 21, 2012
Publication date	Dec 25, 2018
Grant date	Dec 25, 2018

Abstract

An image decoding method includes: receiving information on a set of reference pictures for configuring a set of reference pictures of a current picture, wherein the information on the set of reference pictures includes MSB that may calculate the MSB of POC of a long-term reference picture relative to the current picture, and flag information that represents whether there is MSB; and eliciting the set of reference pictures by using received MSB when the flag information is 1, and performing marking on the reference picture, wherein the flag information may be 1 when a temporal sub-layer identifier is 0, and there is at least one POC value for which a remainder obtained by dividing by a maximum value MaxPicOrderCntLsb is the same as LSB of the POC of the long-term reference picture, in a set of POCs of a previous picture including POC values related to the previous picture.

First claim

What is claimed is:

1. A video decoding method, by a decoding apparatus, the method comprising:
receiving reference picture set (RPS) information for a current picture, wherein the RPS information comprises flag information indicating whether most significant bit (MSB) information for calculating an MSB of a picture order count (POC) of a long-term reference picture (LTRP) for the current picture is present, and wherein the RPS information further comprises the MSB information when a value of the flag information is equal to 1;
deriving the POC of the LTRP based on the MSB information when the value of the flag information is equal to 1,
wherein the value of the flag information is equal to 1 when there is more than one POC value, in a previous picture POC set, for which the POC value modulo MaxPicOrderCntLsb is equal to a least significant bit (LSB) of the POC of the LTRP, MaxPicOrderCntLsb is a predetermined max least significant bit value, wherein the previous picture POC set is derived based on a previous picture of the current picture in decoding order, and the previous picture has a temporal sub-layer identifier equal to 0, and
wherein the previous picture POC set derived based on the previous picture of the current picture includes 1) a POC of each reference picture in an RPS for the previous picture in decoding order, and 2) a POC of each picture following the previous picture in decoding order and preceding the current picture in decoding order;
configuring an RPS for the current picture based on the POC of the LTRP;
performing inter prediction for a block in the current picture based on the RPS for the current picture; and
reconstructing the current picture based on the result of the inter prediction.

2. The method of claim 1 , wherein the previous picture cannot be discarded without affecting decodability of another picture in the same temporal layer, and
wherein the previous picture has a network abstraction layer (NAL) unit type not equal to TRAIL_N, TSA_N, STSA_N, RADL_N, RASL_N, RASL_R, RSV_VCL_N10, RSV_VCL_N12, or RSV_VCL_N14.

3. The method of claim 1 , wherein the previous picture is not a sub-layer non-reference picture which is not used for inter prediction in the decoding process of subsequent pictures of the same sub-layer in decoding order.

4. The method of claim 1 , wherein the previous picture POC set further include a POC of the previous picture.

5. The method of claim 1 , wherein the MSB information is MSB cycle information determining a value corresponding to a difference in POC MSB between the current picture and the LTRP.

6. The method of claim 1 , further comprising constructing a POC list for deriving the RPS,
wherein the POC list comprises a short-term reference picture (STRP) POC list comprising a POC of an STRP for the current picture and an LTRP POC list comprising the POC of the LTRP for the current picture, the STRP POC list is generated using a difference in POC between the current picture and the STRP, and the LTRP POC list is generated by an operation that is POC of current picture−DeltaPocMsbCycleLt*MaxPicOrderCntLsb−POC LSB of current picture+POC LSB of LTRP, DeltaPocMsbCycleLt having a value corresponding to a difference in POC MSB between the current picture and the LTRP.

7. The method of claim 6 , wherein the deriving the RPS comprises:
constructing an LTRP set using a picture with a POC LSB or POC equal to a POC in the LTRP POC list when there is the picture in a decoded picture buffer (DPB) storing a reconstructed picture;
marking all pictures in the LTRP set as “used for long-term reference;”
constructing an STRP set using a picture with a POC equal to a POC in the STRP POC list when there is the picture in the DBP; and
marking all reference pictures in the DPB which are not included on the LTRP set and the STRP set as “unused for reference”.

8. The method of claim 1 , further comprising:
receiving slice type information on the current picture and prediction mode information on the current picture,
wherein the reconstructing the current picture comprises:
determining based on the prediction mode information whether a prediction target block is subjected to intra prediction or inter prediction;
deriving a reference picture list based on the RPS when inter prediction is applied to the prediction target block;
deriving a predictive sample value of the prediction target block using the reference picture list; and
generating a reconstructed picture based on the predictive sample value of the prediction target block.

9. The method of claim 8 , wherein the reconstructed picture is marked as “used for short-term reference”.

10. A decoding apparatus for video decoding, the apparatus comprising:
a decoder configured to receive reference picture set (RPS) information f for a current picture, wherein the RPS information comprises flag information indicating whether most significant bit (MSB) information for calculating an MSB of a picture order count (POC) of a long-term reference picture (LTRP) for the current picture is present and wherein the RPS information further comprises the MSB information when a value of the flag information is equal to 1, to derive the POC of the LTRP based on the MSB information when the value of the flag information is equal to 1, to configure an RPS for the current picture based on the POC of the LTRP, to perform inter prediction for a block in the current picture based on the RPS for the current picture, and to reconstruct the current picture based on the result of the inter prediction; and
a memory configured to store the reconstructed current picture,
wherein the value of the flag information is equal to 1 when there is more than one POC value, in a previous picture POC set, for which the POC value modulo MaxPicOrderCntLsb is equal to a least significant bit (LSB) of the POC of the LTRP, MaxPicOrderCntLsb is a predetermined max least significant bit value, wherein the previous picture POC set is derived based on a previous picture of the current picture in decoding order, and the previous picture has a temporal sub-layer identifier equal to 0, and
wherein the previous picture POC set derived based on the previous picture of the current picture includes 1) a POC of each reference picture in an RPS for the previous picture in decoding order, and 2) a POC of each picture following the previous picture in decoding order and preceding the current picture in decoding order.

11. The apparatus of claim 10 , wherein the previous picture cannot be discarded without affecting decodability of another picture in the same temporal layer, and
wherein the previous picture has a network abstraction layer (NAL) unit type not equal to TRAIL_N, TSA_N, STSA_N, RADL_N, RASL_N, RASL_R, RSV_VCL_N10, RSV_VCL_N12, or RSV_VCL_N14.

12. The apparatus of claim 10 , wherein the previous picture is not a sub-layer non-reference picture which is not used for inter prediction in the decoding process of subsequent pictures of the same sub-layer in decoding order.

13. The apparatus of claim 10 , wherein the previous picture POC set further include a POC of the previous picture.

14. The apparatus of claim 10 , wherein the MSB information is MSB cycle information determining a value corresponding to a difference in POC MSB between the current picture and the LTRP.

15. The apparatus of claim 10 , wherein the decoder constructs a POC list for deriving the RPS,
wherein the POC list comprises a short-term reference picture (STRP) POC list comprising a POC of an STRP for the current picture and an LTRP POC list comprising the POC of the LTRP for the current picture,

Assignees

• Lg Electronics Inc

Inventors

• Hendry Hendry
• Park Joonyoung
• Kim Chulkeun
• Jeon Byeongmoon
• Kim Jungsun

Classifications

• H04N19/46

  Embedding additional information in the video signal during the compression process (H04N19/517, H04N19/68, H04N19/70 take precedence) · CPC title

• H04N19/184

  the unit being bits, e.g. of the compressed video stream · CPC title

• H04N19/423 Primary

  characterised by memory arrangements (H04N19/433 takes precedence) · CPC title

• H04N19/105

  Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction · CPC title

• H04N19/70 Primary

  characterised by syntax aspects related to video coding, e.g. related to compression standards · CPC title