Image encoding method, image decoding method, image encoding device, and image decoding device
An image encoding method for encoding a multiview video is provided. The method includes: determining a maximum number of per-picture pixels and a maximum buffer size which corresponds to a maximum number of candidate reference images for use in a non-multiview coding, based on a level signal representing a coding level with reference to a table; calculating a maximum number of candidate reference views for use in inter-view predictive coding using the maximum number of per-picture pixels, an image size of an input image, and a scale factor for use in multiview video coding; and calculating an MVC maximum buffer size corresponding to a maximum number of candidate reference images for use in multiview video coding, using the maximum number of views, and the maximum buffer size.
1. An image encoding method for encoding a multiview video, comprising:
determining a maximum number of per-picture pixels and a maximum number of candidate reference images which is used in non-multiview coding, from a level signal indicating a coding level with reference to a table indicating a relationship between (i) the coding level, and (ii-i) a maximum number of candidate reference images for non-multiview coding and (ii-ii) a maximum number of per-picture pixels, the maximum number of per-picture pixels indicating a maximum number of pixels per picture and being processable by an image encoding device and an image decoding device;
calculating a maximum number of candidate reference views for inter-view predictive coding, using the maximum number of per-picture pixels, an image size of an input image, and a scale factor for multiview video coding; calculating a maximum number of candidate reference images for multiview video coding, using the maximum number of candidate reference views and the maximum number of candidate reference images for non-multiview coding; and
encoding the multiview video according to the calculated maximum number of candidate reference views and the calculated maximum number of candidate reference images;
wherein in the calculating of a maximum number of candidate reference views, the maximum number of candidate reference views is calculated according to a mathematical expression below:
MaxView=Floor(mvcScaleFactor*MaxLumaFS/(PicHeight*PicWidth)),
where MaxView represents the candidate reference view,
mvcScaleFactor represents the scale factor, and
MaxLumaFs represents the maximum number of per-picture pixels.
2. The image encoding method according to claim 1 ,
wherein in the calculating of a maximum number of candidate reference images for multiview video coding, the maximum number of candidate reference images for multiview video coding is calculated according to a mathematical expression below:
MvcMaxDPBSize=MaxView*MaxDPBSize,
where MvcMaxDPBSize represents the maximum number of candidate reference images for multiview video coding, and
MaxDPBSize represents the maximum number of candidate reference images for non-multiview coding.
3. The image encoding method according to claim 1 , further comprising
setting the number of candidate reference images for multiview video coding to be stored in a decoded picture buffer within a range not exceeding the maximum number of candidate reference images for multiview video coding.
4. An image encoding device for encoding a multiview video, comprising:
processing circuitry;
storage accessible from the processing circuitry,
wherein the processing circuitry executes, using the storage, the following steps:
determining a maximum number of per-picture pixels and a maximum number of candidate reference images which is used in non-multiview coding, from a level signal indicating a coding level with reference to a table indicating a relationship between (i) the coding level, and (ii-i) a maximum number of candidate reference images for non-multiview coding and (ii-ii) a maximum number of per-picture pixels, the maximum number of per-picture pixels indicating a maximum number of pixels per picture and being processable by an image encoding device and an image decoding device;
calculating a maximum number of candidate reference views for inter-view predictive coding, using the maximum number of per-picture pixels, an image size of an input image, and a scale factor for multiview video coding;
calculating a maximum number of candidate reference images for multiview video coding, using the maximum number of candidate reference views and the maximum number of candidate reference images for non-multiview coding; and
encoding the multiview video according to the calculated maximum number of candidate reference views and the calculated maximum number of candidate reference images,
wherein in the calculating of a maximum number of candidate reference views, the maximum number of candidate reference views is calculated according to a mathematical expression below:
MaxView=Floor(mvcScaleFactor*MaxLumaFS/(PicHeight*PicWidth)),
where MaxView represents the candidate reference view,
mvcScaleFactor represents the scale factor, and
MaxLumaFs represents the maximum number of per-picture pixels.
5. An image encoding method for encoding a multiview video, comprising:
determining a maximum number of per-picture pixels and a maximum number of candidate reference images which is used in non-multiview coding, from a level signal indicating a coding level with reference to a table indicating a relationship between (i) the coding level, and (ii-i) a maximum number of candidate reference images for non-multiview coding and (ii-ii) a maximum number of per-picture pixels, the maximum number of per-picture pixels indicating a maximum number of pixels per picture and being processable by an image encoding device and an image decoding device;
calculating a maximum number of candidate reference views for inter-view predictive coding, using the maximum number of per-picture pixels, an image size of an input image, and a scale factor for multiview video coding; calculating a maximum number of candidate reference images for multiview video coding, using the maximum number of candidate reference views and the maximum number of candidate reference images for non-multiview coding;
setting the number of candidate reference images for multiview video coding to be stored in a decoded picture buffer within a range not exceeding the maximum number of candidate reference images for multiview video coding; and
encoding the multiview video according to the calculated maximum number of candidate reference views and the calculated maximum number of candidate reference images;
wherein in the calculating of a maximum number of candidate reference views, the maximum number of candidate reference views is calculated according to a mathematical expression below:
MaxView=Floor(mvcScaleFactor*MaxLumaFS/(PicHeight*PicWidth)),
where MaxView represents the candidate reference view,
mvcScaleFactor represents the scale factor, and
MaxLumaFs represents the maximum number of per-picture pixels.
6. The image encoding method according to claim 5 ,
wherein in the calculating of a maximum number of candidate reference images for multiview video coding, the maximum number of candidate reference images for multiview video coding is calculated according to a mathematical expression below:
MvcMaxDPBSize=MaxView*MaxDPBSize,
where MvcMaxDPBSize represents the maximum number of candidate reference images for multiview video coding, and
MaxDPBSize represents the maximum number of candidate reference images for non-multiview coding.
7. The image encoding method according to claim 5 , further comprising
setting the number of candidate reference images for multiview video coding to be stored in a decoded picture buffer within a range not exceeding the maximum number of candidate reference images for multiview video coding.