[DESCRIPTION] [Title of Invention] CODING METHOD, DECODING METHOD, CODING APPARATUS, DECODING APPARATUS, PROGRAM, AND INTEGRATED CIRCUIT [Technical Field] [0001] The present invention can be used as any method of coding and decoding multimedia data and, more particularly, as a method of coding and decoding H.264/MPEG-4 AVC Multiview Video. [Background Art] [0002] 3D video has been around for a long time. There are several methods of producing the visual effects of 3D to a viewer. One of the methods is to present two sets of moving pictures separately to the viewer's left and right eyes. This is called stereoscopic imaging which involves the capturing of the two sets of moving pictures using two cameras. One of the techniques used previously to present stereoscopic video is using the filtering of the color components where each eye can only view. Such techniques reduce the resolution of a picture that reaches each eye. [0003] Recently, with the advancement in display technologies, viewers are now able to view full resolution videos with each eye. The video standard H.264/MPEG-4 AVC Multiview Video Coding (MVC) is designed for the compression of such 3D imaging where each view is presented in the full resolution. [0004] The video standard H.264/MPEG-4 AVC Multiview Video Coding (MVC) provides a set of compression tools which allow the efficient compression of moving pictures targeted for more than one set of views. The video standard MVC allows the compression of pictures using predictive coding from reconstructed pictures of a different set of views. This "inter-view" prediction utilizes the correlation of pictures captured by different cameras at approximately the same time to efficiently compress the pictures. [0005] In the video standard MVC, the "inter-view" prediction is performed only on pictures of different views having the same picture order count information. The picture order count information is used to indicate the order of reconstructed pictures of the same view. In the video standard MVC, pictures (or view components as defined in the MVC standard) of different views having the same picture order count information are grouped together as a container called an access unit where the size of these access units are constrained by limits defined by a buffer model specified in the video standard. Such constraints are necessary to ensure correct decoding of the compressed video. A view component as defined in the MVC standard as well as in DESCRIPTION is a coded representation of a view in a single access unit. One example of a view is a picture. [0006] The H.264/MPEG-4 AVC High Profile is widely used in various applications, such as High Definition Storage Media and High Definition Digital Video Broadcasting. The Multiview High Profile defined in the video standard MVC is designed as an extension of the H.264/MPEG-4 AVC High Profile where existing implementation of H.264/MPEG-4 AVC High Profile decoders can be easily modified to support the decoding of compressed video streams using the Multiview High Profile. In some implementation of the MVC standard, the decoding of each view can be achieved independently by using the implemented H.264/MPEG-4 AVC decoders. [Summary of Invention] [Technical Problem] [0007] Video content can be captured by a progressive scan or interlaced scan camera. In the case when the video is captured using the interlaced scan camera, the H.264/MPEG-4 AVC High Profile provides a set of coding tools targeted specifically to handle interlaced scan captured moving pictures. Each picture can be coded as a frame or fields as shown in FIGS 1A and 1B. FIG. 1A shows pixel sampling locations of a frame picture including two of the interlaced scanned fields. FIG. 1B shows pixel sampling locations of field pictures where each interlaced scanned field is coded as a field picture. In the case when the two complementary fields are coded as a frame as shown in FIG. 1A, the frame has two picture order counts to present the order of each field. [0008] One problem with the prior art (Multiview Video Coding Standard) is that the definition of the container, or access unit, when each of the pictures (or view components as defined In the MVC standard) of different views is coded differently either using frame coding or field coding. FIG. 2 shows an example of access units with different picture coding structures for each view. As shown in FIG. 2, an access unit container A as shown in a module 200 contains one frame picture of a view 2 and one field picture of a view 1 whereas an access unit container B as shown in a odule 202 contains only one field picture of the view 1. This reduces the compression efficiency because the frame picture of the view 2 in the access unit A cannot fully utilize the inter-view prediction from the view 1. [0009] The problem is that each access unit container does not always contain all the pictures (or view components) of different views. Thus, it is not possible to constrain the size of each access unit by the limits defined by a buffer model specified in the video standard to ensure proper decoding of the compressed video. [0010] The other problem with the prior art (Multiview Video Coding Standard) is the constraints set to restrict the size of the access unit container which contains pictures (or view components as defined in the MVC standard) of more than one views. Even though the maximum size of an access unit is restricted, the maximum size of each picture (or view components as defined in the MVC standard) of each view within the access unit container is not constrained. This will create problems for some implementation of the MVC decoder where each view is decoded by a separate H.264/MPEG-4 AVC decoder. [0011] The constraints defined to restrict the maximum size of an access unit does not ensure the proper decoding of each view component within the access unit when each view is separately decoded by an independent decoder. Such constraints include the maximum number of slices and the size of each picture (or view component as defined in the MVC standard) within an access unit. [0012] The present invention has been conceived in view of the circumstances, and has an object of providing (i) a coding method and (ii) a decoding method for improving the coding efficiency of a group of pictures having different views, and (iii) a coding apparatus, (iv) a decoding apparatus, (v) a program, and (vi) an integrated circuit for implementing the coding method and the decoding method. [Solution to Problem] [0013] The coding method according to an aspect of the present invention is a coding method of coding a first picture group and a second picture group that are interlaced and captured at different view points, and the coding method includes: defining an access unit including a picture of the first picture group and a picture of the second picture group that corresponds to the picture of the first picture group; and coding each of the pictures included in the access unit, the access unit being defined in the defining, wherein the defining includes: determining a unit of coding for determining whether the pictures included in the access unit are to be uniformly coded on a per-field basis or on a per-frame basis; and determining a field type for determining whether the pictures are to be uniformly coded as top fields or bottom fields, when it is determined in the determining of a unit of coding that the pictures included in the access unit are to be coded on a per-field basis, and in the coding, each of the pictures is coded for each access unit in a format determined in the determining of a unit of coding and in the determining of a field type. [0014] According to the method, since the format (field/frame, top/bottom) of all the pictures included in the same access unit is unified, the coding efficiency is improved when the inter-view coding is performed. [0015] Furthermore, in the determining of a unit of coding, a value indicating whether a picture is on a per-field basis or on a per-frame basis may be set to a reference field_pic_flag, in the determining of a field type, a value indicating whether the picture is a top field or a bottom field may be set to a reference bottom_field_flag, and in the coding: the picture of the first picture group and the picture of the second picture group that corresponds to the picture of the first picture group may be obtained based on the values set in the reference field_pic_flag and the reference bottom_field_flag; the value of the reference field_pic_flag may be set to a field_pic_flag included in a slice header of each of the obtained pictures; and the value of the reference bottom_field_flag may be set to a bottom_field_flag included in the slice header, when the value indicating that the picture is on a per-field basis is set to the reference field_pic_fiag. [0016] Furthermore, the coding may include: coding the picture of the first picture group, using a predictive picture generated from a picture of the first picture group; and coding the picture of the second picture group, using one of (i) a predictive picture generated from a picture of the second picture group and (ii) the predictive picture generated from the picture of the first picture group, within the access unit including the picture of the second picture group. Since the format of all the pictures included in the same access unit is unified, the coding efficiency is improved when the inter-view predictive coding is performed in coding of the picture of the second picture group. [0017] The coding method may further include checl