Video decoding apparatus and method

This is a continuation of application Ser. No. 10/833,980, filed Apr. 27, 2004, which claims the benefit of priority to Japanese Application No. 2003-123977, filed on Apr. 28, 2003, in the Japanese Patent Office. The entire contents of both applications are incorporated herein by reference.

1. Field of the Invention

This invention relates to a video decoding apparatus and method, and more particularly, is suitably applied to a video decoding apparatus, for example, in conformity with the Joint Model of Enhanced-Compression Video Coding (JVT) scheme.

2. Description of the Related Art

Video processing apparatuses in conformity with a video coding scheme such as Moving Picture Experts Group (MPEG) for efficient information transmission and storage have been popular for both information providers such as broadcast stations and information receiver such as general users.

Specifically, the MPEG2 (ISO/IEC 13818-2) format is defined as a general-purpose video coding scheme, and now it is widely used for various applications for professionals and consumers because it can handle all of an interlace scanning (interlace) format, a progressive scanning (non-interlace) format, a standard-resolution image format, and a high-resolution image format.

With this MPEG2 format, standard resolution (720×480 pixels) images of the interlace format and high resolution (1920×1088 pixels) images of the interlace format can be transmitted at bit rates of 4-8 Mbps and 18-22 Mbps, respectively.

With popularization of portable terminals such as mobile telephones, a coding scheme at a much higher compression rate is demanded. To meet this demand, an MPEG4 format was approved as a new video coding scheme as ISO/IEC 14496-2 in December 1998.

Further, for video coding for video conference, the standardization of a video coding scheme called H. 26L (ITU-T Q6/16 VCEG) has been progressing. This H. 26L is known as a coding scheme providing higher coding efficiency although more operations are required for coding and decoding, as compared with conventional coding schemes such as the MPEG2 and MPEG4.

In addition, the standardization of a JVT coding scheme which is a video coding scheme providing much higher coding efficiency have been progressing as well. This JVT coding scheme is realized based on the H. 26L by also employing functions that the H. 26L does not have (for example, refer to non patent reference DRAFT ISO/IEC 1/4 496-10:2002 (E)).

By the way, if image data (hereinafter, referred to as JVT coded image data) subjected to compression and coding with the JVT coding scheme as described above can be decoded by only reproducing (decoding) the image data of I-pictures (intra-coded picture) 1_I out of I-pictures, P-pictures (inter frame predictive-coded picture) 1_p and B-pictures (bidirectionaly predictive-coded picture) 1_B, as shown in FIG. 1, a motion compensation process which is required for the reproduction of the P-pictures 1_p and B-pictures 1_B can be omitted, thereby realizing high-speed reproduction.

The JVT coding scheme, however, defines that a filtering process (hereinafter, referred to as a deblock filtering process) is performed on block borders of decoded pictures in order to reduce block noise specific to a block segmentation coding format. Although the existing techniques perform a reproduction process of only I-pictures 1_I, they cannot execute sufficient high-speed reproduction because the deblock filtering process requires a large amount of operations.

As a coding mode in a case where pictures to be coded is in the interlace format, the JVT coding scheme has a picture-based coding mode with a frame as shown in FIG. 2A or a field as shown in FIG. 2B (first and second fields) as a coding unit and a macroblock-based coding mode with a macroblock pair 3 composed of two macroblocks 2 locating above and below as shown in FIG. 3 as a coding unit.

Note that, in the field-structured coding mode with a field as shown in FIG. 2B as a coding unit, the value of field_pic_flag included in a slice header of JVT coded image data is set to “1”. In the frame-structured coding mode with a frame as shown in FIG. 2A as a coding unit, the value of field_pic_flag is set to “0” and the value of mb_adptive_frame_field_flag included in a sequence parameter set is set to “0”. In the coding mode with the macroblock pair 3 as a coding unit as shown in FIG. 3, the value of field_pic_flag of a slice header is set to “0” and the value of mb_adptive frame_field_flag of a sequence parameter set is set to “0”.

In addition, in the JVT coding scheme, considering that the field-structured coding is more effective for pictures of the interlace format having a large amount of motion, than the frame-structured coding, the frame-structured coding and the field-structured coding are adaptively changed picture by picture when pictures to be coded are in the interlace format.

Further, when the field structure is applied as a coding unit in a case where pictures to be coded are in the interlace format, the JVT coding scheme allows the first field of an I-picture 1_I to be coded within this field (hereinafter, referred to as intra-field coding) (I-field) and the second field to be coded by reference to past field pictures (P-field) as shown in FIG. 4.

Therefore, to decode I-pictures I₁, of which the first and second fields has been coded as an I-field and a P-field, respectively, corresponding reference pictures have to be decoded in advance for decoding the second field (P-field).

For this case, in the JVT coding scheme, a multiple reference frame function is supported to use more than two frame pictures before and after a target frame picture as reference pictures for the motion compensation process at the time of coding as shown in FIG. 5.

Therefore, as shown in FIG. 6, to decode the I-picture of which the second field is a P-field, if the P-field uses field pictures of P-pictures 1_p and B-pictures I_B other than the I-picture as reference pictures, the P-pictures I_p and B-pictures I_B should be decoded.

In short, in the JVT coding scheme, in a case where pictures to be coded are in the interlace format, the reproduction process of only I-pictures 1_I may not create decoded video, so that this scheme has a problem that high-speed reproduction by reproduction of only I-pictures 1_I can not be realized.