JP5341786B2 - Image coding apparatus and image conversion apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology capable of preventing image quality from deteriorating when encoding non-compressed image data. <P>SOLUTION: This image encoding device 1 divides input non-compressed image data into a plurality of macro blocks (MB). A peripheral block specifying part 111 specifies an encoding object MB and specifies peripheral MBs from among the encoded macro blocks. When the encoding object MBs constitute a macro block pair, a macro block which does not constitute the macro block pair with the encoding object MB belonging thereto is specified as the peripheral MB. When a peripheral MB constitutes the macro block pair, an encoding object MB and the peripheral MB are coincident in a field classification. If the number of intra-encoded peripheral MBs is smaller than a threshold, an abbreviation mode permitting part 112 permits the application of a skip mode or a direct mode for the encoding object MBs. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to MPEG or H.264. The present invention relates to a technique for encoding image data based on a standard such as H.264.

  An image distributed by digital broadcasting, an image stored in a DVD (Digital Versatile Disk), a hard disk device, or the like is compressed according to various encoding methods. Image compression is performed for the purpose of suppressing transmission band compression, improving transmission speed, and reducing storage size.

  MPEG2 is an encoding method that has been conventionally employed. MPEG2 is an encoding method that can be used not only for recording on a recording medium but also for communication media and broadcast media. Specifically, MPEG2 is widely used for digital broadcasting, video conferencing, videophone systems, and the like.

  As an encoding method, a standard different from MPEG2 is H.264. H.264. H. H.264 is improved in compression rate by about 2 to 4 times compared with MPEG2. Since there are a plurality of different encoding methods, the encoding method may be converted for the purpose of reducing the amount of encoded image data.

JP 2008-206151 A

  H. In H.264 or MPEG2, a macroblock type is determined for each of a plurality of macroblocks of uncompressed image data. Each macroblock is encoded based on the determined macroblock type. Macro block types include special modes called direct mode (H.264 only) and skip mode.

  In the skip mode, motion information and pixel information of the macroblock are not encoded. In the direct mode, macroblock motion information is not encoded. By using the skip mode or the direct mode, the compression rate of the image data can be improved.

  The encoding method disclosed in Patent Document 1 discloses a criterion for determining whether to apply a skip mode to a macroblock. In the encoding method according to Patent Document 1, when the coefficient block of the quantized macroblock is smaller than a predetermined threshold, the skip mode is applied to the quantized macroblock.

  H. When reproducing H.264 image data, a motion vector of a macro block to which a skip mode is applied (a skip macro block) is predicted. When all the macroblocks around the skip macroblock are intra-coded, the size of the motion vector of the skip macroblock is zero. For displaying the skip macroblock, pixel information of the reference image corresponding to the position of the skip macroblock is used.

  When the motion vector of the skip macroblock becomes 0, the subjective image quality may be deteriorated. In a scene with a large motion, the difference between the reference image and the encoding target image is large, so that the macroblock is easily intra-coded. Consider a case where a skip macroblock exists among a plurality of intracoded macroblocks in a scene with large motion. In this case, there is no motion in the skip macroblock and motion occurs around the skip macroblock, so that the viewer feels uncomfortable with the motion of the reproduced image data. Thus, there is a problem that the subjective image quality of the reproduced image is deteriorated by using the skip mode for encoding the non-compressed image data.

  In view of the above problems, an object of the present invention is to provide a technique capable of preventing deterioration in image quality when encoding uncompressed image data.

In order to solve the above-mentioned problem, the invention according to claim 1 is an image encoding device, wherein uncompressed image data is divided into a plurality of blocks, and a block type determination for determining a block type to be applied to each block is determined. And an encoding unit that encodes each block based on the determined block type, wherein the block type determination unit is already encoded by the encoding target block and the encoding unit, and the encoding is performed. A block specifying unit that specifies peripheral blocks located around the target block from the plurality of blocks, and an abbreviated mode in which encoding of a part of the information of the encoding target block is omitted. whether it applied as a type, and a skip mode permission unit determines on the basis of the block information related to the coding of the neighboring blocks, the omission mode Authorization unit, the number of intra-coded the neighboring blocks is less than a predetermined number, and permits the application of the optional mode as the block type of the coding target block.

According to a second aspect of the invention, in the image encoding apparatus according to claim 1, wherein the encoding unit is a block pair units constituted by two blocks adjacent to each other in the vertical direction, interlaced uncompressed image data When the reference pair, which is a block pair including the encoding target block, is field-encoded, the block specifying unit determines a block pair that is different from the reference pair and is frame-encoded. A block to be configured is specified as the peripheral block.

According to a third aspect of the invention, in the image encoding apparatus according to claim 1, wherein the encoding unit is a block pair units constituted by two blocks adjacent to each other in the vertical direction, interlaced uncompressed image data When the reference pair that is the block pair including the coding target block is field-encoded, the block specifying unit is different from the reference pair and is first-field-encoded field-encoded. And the block having the same field type as the encoding target block is specified as the peripheral block among the two blocks constituting the first block pair.

According to a fourth aspect of the present invention, in the image encoding device according to any one of the first to third aspects, the skip mode is a skip in which encoding of motion information and pixel information of the encoding target block is omitted. Mode.

According to a fifth aspect of the present invention, in the image encoding device according to any one of the first to third aspects, the omission mode is a direct mode in which encoding of motion information of the encoding target block is omitted. It is characterized by including.

According to a sixth aspect of the present invention, in the image coding device according to any one of the first to fifth aspects, the encoding target block is composed of a plurality of sub-blocks, and the abbreviated mode includes each sub-block. A mode in which encoding of motion information is omitted.

The invention according to claim 7 is an image conversion apparatus, an image decoding apparatus for decoding the first encoded image data to generate uncompressed image data, and a second encoding by encoding the uncompressed image data. An image encoding device that generates image data; and a block information acquisition device that acquires block information related to encoding of a plurality of first encoded blocks constituting the first encoded image data, the image encoding An apparatus divides the uncompressed image data into a plurality of blocks, determines a block type to be applied to each block, and encodes each block based on the determined block type And the encoding unit identifies, from the plurality of blocks, an encoding target block and a peripheral block located around the encoding target block, and the peripheral block A block specifying unit that specifies a corresponding position-corresponding block from the plurality of first encoding blocks, and an abbreviated mode in which encoding of a part of the information of the encoding target block is omitted, the block of the encoding target block whether to permit the type, look including a Truncating mode permission unit determines on the basis of the block information of the position corresponding block, the optional mode permission unit is included in the position corresponding block and intra-coded If the number of the first encoded blocks is smaller than a predetermined number, the omission mode is permitted as a block type of the encoding target block .

An eighth aspect of the present invention is the image conversion apparatus according to the seventh aspect , wherein the encoding unit interlaces the uncompressed image data in units of block pairs formed by two blocks adjacent to each other in the vertical direction. When the reference pair that is the block pair including the encoding target block is field-encoded, the block specifying unit configures a block pair that is different from the reference pair and is frame-encoded. A block is specified as the peripheral block.

According to a ninth aspect of the present invention, in the image conversion device according to the seventh aspect , the encoding unit interlaces the uncompressed image data in units of block pairs composed of two blocks adjacent to each other in the vertical direction. When the reference pair, which is a block pair including the encoding target block, is field-encoded, the block specifying unit determines a first block pair that is different from the reference pair and is field-encoded. And selecting a block whose field type matches that of the encoding target block from the two blocks constituting the first block pair as the peripheral block.

According to a tenth aspect of the present invention, in the image conversion device according to the ninth aspect , the block specifying unit, when the position corresponding block is field-encoded and includes a first field and a second field, The first encoding block corresponding to two blocks constituting the block pair and including the position corresponding block is specified, and the abbreviated mode permission unit includes at least one of the two first encoding blocks. If it is intra-coded, it is determined that the position corresponding block is intra-coded.

According to an eleventh aspect of the present invention, in the image conversion device according to any one of the seventh to tenth aspects, the skip mode is a skip in which encoding of motion information and pixel information of the encoding target block is omitted. Mode.

According to a twelfth aspect of the present invention, in the image conversion device according to any one of the seventh to tenth aspects, the omission mode is a direct mode in which encoding of motion information of the encoding target block is omitted. It is characterized by including.

According to a thirteenth aspect of the present invention, in the image conversion device according to any one of the seventh to tenth aspects, the encoding target block is composed of a plurality of sub-blocks, and the abbreviated mode is set for each sub-block. A mode in which encoding of motion information is omitted.

  The image encoding apparatus according to the present invention determines whether or not the omission mode can be applied to the encoding target block based on the block information of the peripheral blocks that have already been encoded. For example, when the number of intra-coded peripheral blocks is smaller than a predetermined number, application of the abbreviated mode to the encoding target block is permitted. Since the application of the omission mode to the encoding target block is suppressed according to the encoding status of the peripheral blocks, the image quality of the encoded image data can be improved.

  The image conversion apparatus according to the present invention specifies an encoding target block and a peripheral block, and specifies a block of first encoded image data corresponding to the peripheral block as a position corresponding block. The image conversion apparatus according to the present invention determines whether or not the omission mode can be applied to the encoding target block based on the block information of the position corresponding block. For example, if the number of intra-corresponding position-corresponding blocks is smaller than a predetermined number, application of the omission mode to the encoding target block is permitted. Alternatively, it is determined whether or not the omission mode can be applied to the encoding target block based on the average of the motion vectors of the position corresponding block. Thus, since the application of the omission mode to the encoding target block is suppressed according to the encoding state of the position corresponding block, the image quality of the second encoded image data can be improved.

It is a block diagram which shows the functional structure of the image coding apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the flow of operation | movement of a macroblock classification determination part. It is a figure which shows the positional relationship of an encoding object macroblock and a periphery macroblock. It is a figure which shows the positional relationship of an encoding object macroblock and a periphery macroblock. It is a block diagram which shows the functional structure of the image conversion apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the flow of operation | movement of a macroblock classification determination part. It is a figure which shows the positional relationship of an encoding object macroblock and a periphery macroblock. It is a figure which shows the positional relationship of an encoding object macroblock and a periphery macroblock. H. 2 is a diagram illustrating a difference in field division between H.264 and MPEG2. It is a block diagram which shows the functional structure of the image conversion apparatus which concerns on the 3rd Embodiment of this invention. It is a figure which shows the positional relationship of an encoding object macroblock and a periphery macroblock.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
{1. overall structure}
FIG. 1 is a block diagram showing a functional configuration of an image encoding device 1 according to the present embodiment. As shown in FIG. 1, the image encoding device 1 converts input uncompressed image data into H.264 format. H.264 encoding and H.264. It is a device that outputs H.264 image data. The image encoding device 1 includes a macroblock type determination unit 11 and an encoding unit 12.

  The macroblock type determination unit 11 determines a macroblock type to be applied to each of a plurality of macroblocks constituting uncompressed image data. Type information related to the determined macroblock type is input to the encoding unit 12. The encoding unit 12 encodes each macroblock based on the determined macroblock type.

  The macroblock type determination unit 11 includes a peripheral block identification unit 111 and an abbreviated mode permission unit 112. The peripheral block specifying unit 111 specifies the encoding target macroblock and the peripheral macroblock. The peripheral macroblock is a macroblock located around the encoding target macroblock. The peripheral macroblock is used to determine the macroblock type of the encoding target macroblock. Hereinafter, the encoding target macroblock is referred to as “encoding target MB (MacroBlock)”. A peripheral macroblock is referred to as a “peripheral MB”.

  The abbreviated mode permission unit 112 determines whether to permit the application of the abbreviated mode to the encoding target MB based on information related to encoding of neighboring MBs. The abbreviated mode refers to a macroblock type that omits encoding of some information related to the macroblock, and corresponds to a skip mode and a direct mode. In the skip mode, pixel information and motion information of the macroblock are not encoded. In the direct mode, macroblock motion information is not encoded.

{2. Flow of operation of macroblock type determination unit 11}
FIG. 2 is a flowchart showing the operation of the macroblock type determination unit 11. In the initial state, application of the omission mode is prohibited as the macroblock type of the encoding target MB.

  Uncompressed image data is input to the image encoding device 1 and divided into a plurality of macroblocks. The macroblock type determination unit 11 determines an encoding target MB from among the plurality of divided macroblocks (step S11).

  The macroblock type determination unit 11 confirms whether there is an encoded macroblock (step S12). When there is no encoded macroblock (No in step S12), the macroblock type determination unit 11 ends the process illustrated in FIG. That is, the application of the omission mode to the encoding target MB is prohibited. This is because it is necessary to specify neighboring MBs when determining whether or not the omission mode can be applied to the encoding target MB.

  On the other hand, when an encoded macroblock exists (Yes in step S12), the peripheral block specifying unit 111 specifies a peripheral MB from the encoded macroblock (step S13). Details of the determination of the peripheral MB (step S13) will be described later. The omitted mode permission unit 112 acquires macroblock information of neighboring MBs from the encoding unit 12. The omitted mode permission unit 112 refers to the macroblock information of the neighboring MB and counts the intra-coded neighboring MB (step S14). The macro block information is information that can specify the encoding mode of each macro block. If the number of intra-encoded neighboring MBs is smaller than the threshold value (Yes in step S15), the omission mode permission unit 112 permits the omission mode to be applied to the encoding target MB (step). S16).

  When the application of the omitted mode is permitted, the macroblock type determination unit 11 can apply the skip mode or the direct mode as the macroblock type of the encoding target MB. However, the skip mode or the direct mode is not always applied to the encoding target MB. This is because the macroblock type determination unit 11 determines the macroblock type to be applied to the encoding target MB based on the encoding cost of the encoding target MB.

  If the number of neighboring MBs that are intra-coded is equal to or greater than the threshold value (No in step S15), the abbreviated mode permission unit 112 ends the process illustrated in FIG. That is, the application of the omission mode to the encoding target MB is prohibited. The threshold value is a numerical value preset in the image encoding device 1. When 1 is set as the threshold value, if there is at least one intra-coded peripheral MB, application of the abbreviated mode to the encoding target MB is prohibited.

{3. Determination of peripheral MB}
When the uncompressed image data is compatible with interlace scanning, the uncompressed image data is divided into two fields, a Top field and a Bottom field. H. In H.264, the two fields are encoded by either frame encoding or field encoding. In frame coding, two fields are coded without distinction. In field coding, the two fields are coded separately.

  H. H.264 includes PAFF (Picture Adaptive Frame-Field Coding) and MBAFF (Macroblock Adaptive Frame-Field Coding) as encoding systems compatible with interlace scanning. In PAFF, frame coding and field coding can be switched in units of pictures. In MBAFF, frame coding and field coding can be switched in units of macroblock pairs. The macroblock pair is used for encoding based on MBAFF, and is composed of two macroblocks adjacent to each other in the vertical direction.

  Depending on whether or not the uncompressed image data is encoded by MBAFF, the specific criteria for neighboring MBs differ. The details of the specific criteria for the peripheral MB will be described separately for a case where uncompressed image data is encoded with MBAFF and a case where it is not encoded with MBAFF.

{3.1. When uncompressed image data is not encoded by MBAFF}
First, a description will be given of a specific criterion for neighboring MBs when uncompressed image data is not encoded by MBAFF. Examples of cases where uncompressed image data is not encoded by MBAFF include encoding based on PAFF or encoding corresponding to progressive scanning.

  FIG. 3 is a diagram showing a positional relationship between the encoding target MB and the neighboring MB when the uncompressed image data is not encoded by MBAFF. The uncompressed image data is composed of 5 × 6 macro blocks m0 to m29. The encoding unit 12 encodes macroblocks m0, m1, m2,..., M28, m29 in this order.

  In the example illustrated in FIG. 3, the macroblocks m0 to m11 are already encoded, and the macroblock m12 is an encoding target MB. Encoded macroblocks m0 to m11 are indicated by hatching or sand. The macroblock indicated by the sand is a macroblock specified as a peripheral MB when the macroblock m12 is the encoding target MB.

  In step S13 (see FIG. 2) described above, the peripheral block specifying unit 111 specifies an encoded macroblock as a peripheral MB. In the present embodiment, peripheral block specifying section 111 specifies four macroblocks m6 to m8, m11 adjacent to macroblock m12 as peripheral MBs.

  The omitted mode permission unit 112 counts the intra-coded macro blocks among the macro blocks m6 to m8 and m11 (step S14). If the count value of the intra-coded macroblock is smaller than the threshold value (Yes in step S15), application of the abbreviated mode to the macroblock m12 is permitted (step S16). Thereby, the direct mode can be applied to the sub-macroblock of the macroblock m12.

  The peripheral block specifying unit 111 may specify macroblocks (macroblocks m4, m6, etc.) that are not adjacent to the encoding target MB as peripheral MBs. Alternatively, the peripheral block specifying unit 111 may specify all encoded macroblocks as peripheral MBs.

{3.2. When uncompressed image data is encoded by MBAFF}
A specific criterion for neighboring MBs when uncompressed image data is encoded by MBAFF will be described. Hereinafter, a case where a macroblock (reference pair) including an encoding target MB is field-encoded and a case where it is frame-encoded will be described.

  First, specific criteria for neighboring MBs when the reference pair is field-encoded will be described. FIG. 4 is a diagram showing the positional relationship between the encoding target MB and the neighboring MB when the reference pair is field-encoded. In FIG. 4, the encoded macroblock is indicated by hatching or sand. The macroblock indicated by the sand is a macroblock specified as a peripheral MB when the macroblock m14 is the encoding target MB. Macroblocks are encoded in the order of macroblocks m0, m1,.

  In FIG. 4, macroblocks m0 to m29 each constitute a macroblock pair. A macroblock pair indicated by a one-dot chain line is frame-encoded. Macroblock pairs P11, P12, and P13 indicated by solid lines are field-encoded. Macroblocks m4 and m5 constitute a macroblock pair P11. Macroblocks m12 and m13 constitute a macroblock pair P12. Macroblocks m14 and m15 constitute a macroblock pair P13. Macroblocks m4, m12, and m14 are Top fields. Macro blocks m5, m13, and m15 are the Bottom field.

The macro block m14 is an encoding target MB, and the macro block pair P13 is a reference pair. When the reference pair is field-encoded, a macroblock that satisfies one of the following two criteria is specified as a neighboring MB.
(1) A macroblock that is different from the reference pair and constitutes a macroblock pair to be frame-encoded.
(2) A macroblock whose field type matches the encoding target MB, out of two macroblocks that are different from the reference pair and constitute a field-encoded macroblock pair.

  In the example illustrated in FIG. 4, the peripheral block specifying unit 111 specifies macroblocks m3 and m7 adjacent to the macroblock m14 as the peripheral MB corresponding to the criterion (1). The peripheral block specifying unit 111 may specify macroblocks m1 and m10 as peripheral blocks in addition to the macroblocks m3 and m7 as long as the criterion (1) is satisfied. Alternatively, the peripheral block specifying unit 111 may specify all the encoded macro blocks that satisfy the criterion (1) as the peripheral MB.

  The peripheral block specifying unit 111 specifies macroblock pairs P11 and P12 adjacent to the reference pair (macroblock pair P13) as the peripheral MB corresponding to the reference (2). Of the macroblocks included in the macroblock pairs P11 and P12, the macroblocks m4 and m12 are specified as the peripheral MB. This is because the field type of the macroblocks m4 and m12 matches the field type of the macroblock m14. If the peripheral block specifying unit 111 satisfies the criterion (2), the peripheral block specifying unit 111 may specify a peripheral MB from macroblocks that constitute a macroblock pair that is not directly adjacent to the reference pair.

  In FIG. 3, when uncompressed image data is not encoded by MBAFF, a macroblock (macroblock m7) immediately adjacent to the encoding target MB (macroblock m12) is specified as a peripheral MB. However, in the example shown in FIG. 4, the macroblock m5 immediately adjacent to the encoding target MB (macroblock m14) is not specified as a peripheral MB of the macroblock m14 because the field type is different from that of the macroblock m14. Thus, the macroblock specified as the peripheral MB varies depending on how the uncompressed image data is encoded.

  Next, specific criteria for neighboring MBs when the reference pair is frame-encoded will be described. In this case, the peripheral block specifying unit 111 specifies the peripheral MB based on the same standard (see FIG. 3) as when uncompressed image data is not encoded by MBAFF. That is, when the reference pair is frame-encoded, the peripheral block specifying unit 111 surrounds the encoded macroblock regardless of whether other macroblock pairs are field-encoded or frame-encoded. It can be specified as MB. This is because since the encoding target MB includes a Top field and a Bottom field, it is not necessary to confirm the type of the encoded macroblock.

  After the neighboring MB is specified in step S13, the omitted mode permission unit 112 determines whether or not the omitted mode can be applied to the encoding target MB based on the macroblock information of the neighboring MB (step S14 to step S16).

  As described above, the macroblock type determination unit 11 permits application of the abbreviated mode to the encoding target MB if the number of intra-coded peripheral MBs is smaller than the threshold value. This suppresses the application of the abbreviated mode to the macroblock when encoding a scene where the macroblock is often intra-encoded (such as a scene with a large amount of motion). H. The viewer does not feel uncomfortable during playback. H.264 image data can be generated.

[Second Embodiment]
{1. overall structure}
FIG. 5 is a block diagram illustrating a functional configuration of the image conversion apparatus 100 according to the present embodiment. The image conversion apparatus 100 converts MPEG2 image data to H.264 format. To H.264 image data. The image conversion apparatus 100 includes an image decoding apparatus 2 and an image encoding apparatus 3.

  The following description will focus on differences from the first embodiment.

  The image decoding device 2 decodes MPEG2 image data to generate uncompressed image data. Uncompressed image data is input to the image encoding device 3. Similar to the image encoding device 1 according to the first embodiment, the image encoding device 3 converts the uncompressed image data to H.264 format. It encodes by H.264 system. The image encoding device 3 includes a macroblock type determination unit 11, an encoding unit 12, and a macroblock information acquisition unit 31.

  The macroblock information acquisition unit 31 acquires macroblock information of each macroblock of MPEG2 image data from the image decoding device 2. The macroblock type determination unit 11 acquires macroblock information of a macroblock of MPEG2 image data corresponding to the neighboring MB from the macroblock information acquisition unit 31. The macroblock type determination unit 11 determines whether or not the omission mode can be applied to the encoding target MB based on the acquired macroblock information.

{2. Flow of operation of macroblock type determination unit 11}
FIG. 6 is a flowchart showing an operation flow of the macroblock type determination unit 11 in the present embodiment. In the initial state, application of the omission mode is prohibited as the macroblock type.

  The uncompressed image data is input to the image encoding device 3 and divided into a plurality of macroblocks. The macroblock type determination unit 11 determines an encoding target MB from among the plurality of divided macroblocks (step S21).

  The peripheral block determination unit 11 specifies the peripheral MB and the position corresponding MB (step S22). The position corresponding MB is a macro block of MPEG2 image data corresponding to the position of the peripheral MB. The omitted mode permission unit 112 acquires the macro block information of the position-corresponding MB from the macro block information acquisition unit 31 (step S23).

  The omitted mode permission unit 112 counts position-corresponding MBs that are intra-coded based on the acquired macroblock information (step S24). If the number of intra-encoded position-corresponding MBs is smaller than the threshold value (Yes in step S25), the omission mode permission unit 112 permits the omission mode to be applied to the encoding target MB ( Step S26).

  If the number of intra-encoded position-corresponding MBs is equal to or greater than the threshold value (No in step S25), the omission mode permission unit 112 does not permit the application of the omission mode to the encoding target MB. The process shown in FIG.

{3. Identifying neighboring MBs and location-compatible MBs}
Next, the identification of neighboring MBs and position-corresponding MBs (step S22) and the counting of intra-coded position-corresponding MBs (step S24) will be described in detail. The processing in step S22 and step S24 changes according to the following image conversion conditions (pattern A to pattern C).

(Pattern A) Uncompressed image data is not encoded by MBAFF.
(Pattern B) When MPEG2 image data does not support interlace scanning and uncompressed image data is encoded by MBAFF.
(Pattern C) When MPEG2 image data corresponds to interlace scanning and uncompressed image data is encoded by MBAFF.

{3.1. Specification of neighboring MB and position corresponding MB in pattern A}
Specifically, the pattern A is applied when uncompressed image data is encoded by PAFF, or when uncompressed image data is encoded corresponding to progressive scanning. In pattern A, the MPEG2 image data scanning method may be interlace scanning or progressive scanning.

  FIG. 7 is a diagram illustrating a positional relationship between the encoding target MB and the peripheral MB when the pattern A is applied. MPEG2 and uncompressed image data are composed of 5 × 6 macroblocks. Macroblocks are encoded in the order of macroblocks m0, m1,.

  In FIG. 7, the surrounding MB when the encoding target MB is the macroblock m12 is indicated by sand. In pattern A, the neighboring block specifying unit 111 sets eight neighboring macroblocks (macroblocks m6 to m8, m11, m13, m16 to m18) as surrounding MBs so as to surround the encoding target MB (macroblock m12). Identify. Unlike the first embodiment, the pattern A does not consider whether or not the macroblock has been encoded when the neighboring MB is specified.

  The peripheral block specifying unit 111 specifies macroblocks of MPEG2 image data corresponding to the macroblocks m6 to m8, m11, m13, and m16 to m18 as position corresponding MBs. The peripheral block specifying unit 111 may specify macroblocks m2, m10, m28 and the like that are not directly adjacent to the macroblock m12 as the peripheral MB. Alternatively, all macroblocks may be specified as neighboring MBs. The peripheral block specifying unit 111 may specify the macro block of the MPEG2 image data corresponding to the position of the macro block m12 as the position corresponding MB.

  The omitted mode permission unit 112 refers to the macroblock information corresponding to the identified position-corresponding MB, and counts intra-coded position-corresponding MBs (step S24). Based on the count value, it is determined whether or not to apply the omission mode to the macroblock m12 (steps S25 and S26).

{3.2. Identification of neighboring MB and reference MB in pattern B}
Pattern B is applied when MPEG2 image data does not support interlace scanning and uncompressed image data is encoded by MBAFF.

  FIG. 8 is a diagram illustrating a positional relationship between the encoding target MB and the neighboring MB when the pattern B is applied. MPEG2 and uncompressed image data are composed of 5 × 6 macroblocks. Macroblocks are encoded in the order of macroblocks m0, m1,.

  In FIG. 8, when the encoding target MB is the macroblock m14, the macroblock specified as the peripheral MB is indicated by sand. In FIG. 8, macroblocks m0 to m29 each constitute a macroblock pair. A macroblock pair indicated by a one-dot chain line is frame-encoded. Macroblock pairs P21, P22, P23, and P24 indicated by solid lines are field-encoded. Macroblocks m4 and m5 constitute a macroblock pair P21. Macroblocks m10 and m11 constitute a macroblock pair P22. Macroblocks m14 and m15 constitute a macroblock pair P23 (reference pair). Macroblocks m16 and m17 constitute a macroblock pair P24. Macro blocks m4, m10, m14, and m16 are Top fields. Macro blocks m5, m11, m15, and m17 are Bottom fields.

  In the pattern B, the specific reference for the neighboring MBs differs depending on whether the reference pair is field-encoded or frame-encoded. First, specific criteria for neighboring MBs when the reference pair is field-encoded will be described.

When the reference pair is field-encoded in the pattern B, the peripheral block specifying unit 111 specifies a macroblock corresponding to one of the following criteria (B1) and (B2) as a peripheral MB.
(B1) Unlike the reference pair, a macroblock constituting a macroblock pair to be frame-encoded.
(B2) Unlike the reference pair, a macroblock whose field type matches the encoding target MB, out of two macroblocks constituting a macroblock pair to be field-encoded.

  In the example illustrated in FIG. 8, the peripheral block specifying unit 111 specifies macroblocks m3, m7, m12, and m13 that correspond to the reference (B1) and are directly adjacent to the macroblock m14 as peripheral MBs. The peripheral block specifying unit 111 specifies macroblocks m4 and m16 constituting the macroblock pairs P21 and P24 that meet the condition of the reference (B2) and are directly adjacent to the reference pair as the peripheral MB. As a result, the macro block of the MPEG2 image data corresponding to the macro blocks m3, m4, m7, m12, m13, m16 is specified as the position corresponding MB.

  The peripheral block specifying unit 111 may specify macroblocks m1, m22, m24, etc. that correspond to the reference (B1) and are not directly adjacent to the reference pair (macroblock pair P23) as peripheral MBs. The peripheral block specifying unit 111 may specify the peripheral MB from the macroblock pair P22 that corresponds to the reference (B2) and is not directly adjacent to the reference pair. In the example shown in FIG. 8, the macroblock m10 can be specified as the peripheral MB. Similarly to the pattern A, the macro block of the MPEG2 image data corresponding to the macro block m14 may be specified as the position corresponding MB.

  Next, a specific criterion for neighboring MBs when the reference pair is frame-encoded in the pattern B will be described. In this case, the peripheral block specifying unit 111 specifies the peripheral MB on the same basis as the pattern A. That is, when the reference pair is frame-encoded in the pattern B, the peripheral block specifying unit 111 specifies the peripheral MB without considering the field types of other macroblocks.

  The omitted mode permission unit 112 refers to the macroblock information corresponding to the identified position-corresponding MB, and counts intra-coded position-corresponding MBs (step S24). Based on the count value, it is determined whether or not to apply the omission mode to the macroblock m14 (steps S25 and S26).

{3.3. Identification of neighboring MB and reference MB in pattern C}
Pattern C is applied when MPEG2 image data corresponds to interlaced scanning and the reference pair is encoded by MBAFF.

  The specification of the peripheral MB and the position correspondence MB in the pattern C is the same as that in the pattern B. In pattern C, the method of counting intra-coded position-corresponding MBs is different from patterns A and B. This is because of field coding in MPEG2 and H.264. This is because the assignment of field types to macroblocks is different from field coding in units of macroblock pairs in H.264. This will be described in detail below.

  FIG. 9 is a diagram showing field divisions of macro blocks m4 and m5 and field divisions of macro blocks m4i and m5i of MPEG2 image data. The macro block m4 is specified as a peripheral MB. The macro blocks m4 and m5 constitute a macro block pair P21. Macroblocks m4i and m5i correspond to the positions of macroblocks m4 and m5. The macro block m4i is an MB corresponding to the position of the macro block m4 identified as the peripheral MB. Since the macro blocks m4i and m5i are encoded in MPEG2, they do not constitute a macro block pair.

  Since the macro block m4 is located above the macro block pair P21, it is encoded as a Top field. Since the macro block m5 is located below the macro block pair P21, it is encoded as a Bottom field. On the other hand, in MPEG2, the upper eight lines of the macro blocks m4i and m5i are encoded as a Top field. The lower 8 lines are encoded as a Bottom field.

  That is, in the pattern C, the macroblock m4 and the macroblock m4i at the same position do not match the field types even though they are at the same position. The macro block m4 and the upper eight lines of the macro blocks m4i and m5i are in a correspondence relationship. For this reason, if any one of the macro blocks m4i and m5i is intra-coded, the omitted mode permission unit 112 determines that the position-corresponding MB (macro block m4i) is intra-coded. On the other hand, if both macroblocks m4i and m5i are inter-coded, it is determined that the position-corresponding MB is inter-coded. The omitted mode permission unit 112 may determine that the position-corresponding MB (macroblock m4i) is intra-coded if both the macroblocks m4i and m5i are intra-coded.

  The omission mode permission unit 112 determines whether or not the omission mode is permitted for the encoding target MB based on the count value of the intra-corresponding position-corresponding MB (steps S25 and S26).

  As described above, the image conversion apparatus 100 according to the present embodiment determines whether or not the omission mode can be applied to the encoding target MB based on the number of position-corresponding MBs that are intra-coded. By determining the magnitude of motion around the encoding target MB from the MPEG2 image data, more macroblocks can be identified as peripheral MBs than in the first embodiment. Therefore, it is possible to improve the accuracy of determining whether to permit the application mode.

[Third Embodiment]
FIG. 10 is a block diagram illustrating a functional configuration of an image conversion apparatus 200 according to the third embodiment. In the image conversion apparatus 200, the macroblock type determination unit 11 newly includes an average value calculation unit 113. The average value calculation unit 113 calculates the vector average of the motion vectors of the position corresponding MB. The omitted mode permission unit 112 determines whether or not the omitted mode is applicable to the encoding target MB based on the calculated vector average. The following description will focus on differences from the first embodiment and the second embodiment.

  With reference to FIG. 6, the flow of determining whether or not the omission mode can be applied to the encoding target MB will be described. Specifying the encoding target MB (see step S21, FIG. 6) is the same as in the second embodiment. The determination of the peripheral MB and the position-corresponding MB (step S22) is basically determined based on one of the patterns A to C described in the second embodiment.

  However, when the image conversion condition corresponds to the pattern C, the specific criterion for the surrounding MB may be different from that in the second embodiment. Specifically, MPEG2 image data corresponds to interlace scanning, uncompressed image data is encoded by MBAFF, and a reference pair is field-encoded. This will be described in detail below.

  FIG. 11 is a diagram showing the positional relationship between the encoding target MB and the peripheral MB when the pattern C is applied and the reference pair is field-encoded in the present embodiment. As in FIG. 8, macroblock pairs P21 to P24 are configured. Macro blocks m4, m10, m14, and m16 are Top fields. Macro blocks m5, m11, m15, and m17 are Bottom fields.

The peripheral block specifying unit 111 specifies a macroblock corresponding to one of the following criteria (C1) and (C2) as a peripheral MB.
(C1) Unlike the reference pair, a macroblock constituting a macroblock pair to be frame-encoded.
(C2) Two macroblocks constituting a macroblock pair to be field-encoded unlike the reference pair.

  The reference (C1) is the same as the reference (B1) according to the second embodiment. As illustrated in FIG. 11, the neighboring block specifying unit 111 specifies macroblocks m3, m7, m12, and m13 that satisfy the criterion (C1) and are directly adjacent to the encoding target MB (macroblock m14) as neighboring MBs. . In the reference (C1), two macroblocks (for example, macroblocks m3 and m4) constituting a macroblock pair to be frame-encoded may not be simultaneously specified as neighboring MBs.

  The criterion (C2) is different from the condition of the criterion (B2) of the second embodiment, and the field type of the encoding target MB is not considered. That is, two macroblocks constituting a macroblock pair are specified as neighboring MBs regardless of the field type of the encoding target MB. As illustrated in FIG. 11, the neighboring block specifying unit 111 performs macroblocks m4, m5, and m16 from macroblock pairs P21 and P22 that are directly adjacent to the reference pair (macroblock pair P23) as macroblocks that satisfy the reference (C2). , M17 are extracted. As a result, the macroblocks m3 to m5, m7, m12, m13, and m16 to 17 are specified as peripheral MBs of the macroblock m14. Macroblocks of MPEG2 image data corresponding to these macroblocks are specified as position corresponding MBs.

  The peripheral block specifying unit 111 may specify macroblocks m2 and m24 that satisfy the criterion (C1) and are not directly adjacent to the encoding target MB (macroblock m14) as the peripheral MB. The peripheral block specifying unit 111 may specify a macroblock m10 that satisfies the criterion (C2) and is not directly adjacent to the reference pair (macroblock pair P23) as a peripheral MB. Similarly to the pattern A, the macro block of the MPEG2 image data corresponding to the macro block m14 may be specified as the position corresponding MB.

  Reference is again made to FIG. After identifying the peripheral MB and the position-corresponding MB in step S22, the macroblock type determining unit 11 acquires macroblock information of the position-corresponding MB from the macroblock information acquiring unit 31 (step S23). The macroblock information includes the motion vector information of the position corresponding MB.

  Instead of the processing in step S24, the average value calculation unit 113 calculates the vector average of the motion vectors of the position corresponding MB using the motion vector information of the position corresponding MB. The vector average is calculated by dividing the sum of motion vectors of position-corresponding MBs by the number of neighboring MBs.

  The omitted mode permission unit 112 checks whether or not the magnitude of the vector average is equal to or greater than a predetermined value instead of the process in step S25. If the size of the vector average is smaller than the predetermined value, the omission mode permission unit 112 permits application of the omission mode to the encoding target MB (step S26). If the magnitude of the vector average is greater than or equal to a predetermined value, the omission mode permission unit 112 does not permit application of the omission mode to the encoding target MB.

  Thus, in this embodiment, whether or not the omission mode is applicable to the encoding target MB is determined based on the vector average of the motion vectors of the position-corresponding MB. By estimating the motion of the image in the peripheral area of the encoding target MB based on the motion vector of the position-corresponding MB, it is possible to reliably prohibit the application of the abbreviated mode in scenes with large motion. For this reason, it is possible to prevent deterioration in subjective image quality.

  In the third embodiment, the example in which the average value calculation unit 113 calculates the vector average as the average of the motion vectors has been described. However, the average value calculation unit 113 may calculate a scalar average of motion vectors as the average of motion vectors. In this case, if the scalar average is equal to or smaller than the predetermined value, the abbreviated mode permission unit 11 may permit the application of the abbreviated mode to the encoding target MB.

  The average value calculation unit 113 may calculate the variation in the angle of the motion vector as the average of the motion vectors. The variation in angle is a value obtained by selecting one vector from a plurality of motion vectors and averaging the difference in angle between the selected vector and another motion vector. The omission mode permission unit 112 may permit application of the omission mode to the encoding target MB if the variation in angle is equal to or less than a predetermined value.

  In the third embodiment, the applicability of the omission mode to the encoding target MB is determined based on the vector average of motion vectors. However, the omission mode permission unit 112 may permit the application of the omission mode to the encoding target MB when there is no zero vector among the motion vectors of the position corresponding MB. Alternatively, when the number of zero vectors present in the motion vector is equal to or smaller than a predetermined number, application of the abbreviated mode to the encoding target MB may be permitted.

1, 3 Image encoding device 2 Image decoding device 11 Macroblock type determination unit 12 Encoding unit 31 Macroblock information acquisition unit 100, 200 Image conversion device 111 Peripheral block specifying unit 112 Omission mode permission unit 113 Average value calculation unit

Claims (13)

A block type determining unit that divides uncompressed image data into a plurality of blocks and determines a block type to be applied to each block;
An encoding unit that encodes each block based on the determined block type;
With
The block type determination unit
A block identification unit that identifies, from the plurality of blocks, an encoding target block and a peripheral block that is already encoded by the encoding unit and is positioned around the encoding target block;
It is determined based on block information related to encoding of the neighboring blocks whether or not to apply an abbreviated mode in which encoding of a part of information of the encoding target block is omitted as a block type of the encoding target block. An abbreviated mode permission section;
Equipped with a,
The abbreviation mode permission unit
The image coding apparatus according to claim 1, wherein if the number of intra-encoded peripheral blocks is smaller than a predetermined number, application of the abbreviated mode is permitted as a block type of the block to be encoded. The image encoding device according to claim 1 ,
The encoding unit encodes uncompressed image data in correspondence with interlaced scanning in units of block pairs composed of two blocks adjacent to each other in the vertical direction,
When the reference pair which is a block pair including the encoding target block is field-encoded, the block specifying unit is different from the reference pair and a block constituting a frame-encoded block pair is used as the peripheral block. An image encoding device characterized by specifying. The image encoding device according to claim 1 ,
The encoding unit encodes uncompressed image data corresponding to an interlace in units of a block pair constituted by two blocks adjacent to each other in the vertical direction,
The block specifying unit selects a first block pair that is different from the reference pair and is field-encoded when a reference pair that is a block pair including the encoding target block is field-encoded, and An image coding apparatus characterized by specifying, as the peripheral block, a block whose field type matches the coding target block among two blocks constituting a block pair. The image encoding device according to any one of claims 1 to 3 ,
The abbreviated mode is
Skip mode in which encoding of motion information and pixel information of the encoding target block is omitted,
An image encoding device comprising: The image encoding device according to any one of claims 1 to 3 ,
The abbreviated mode is
Direct mode in which encoding of motion information of the encoding target block is omitted,
An image encoding device comprising: The image encoding device according to any one of claims 1 to 5 ,
The encoding target block is composed of a plurality of sub-blocks,
The abbreviated mode is
A mode in which encoding of motion information of each sub-block is omitted,
An image encoding device comprising: An image decoding device that decodes the first encoded image data to generate uncompressed image data;
An image encoding device that encodes uncompressed image data to generate second encoded image data;
A block information acquisition device for acquiring block information relating to encoding of a plurality of first encoded blocks constituting the first encoded image data;
With
The image encoding device includes:
A type determining unit that divides the uncompressed image data into a plurality of blocks and determines a block type to be applied to each block;
An encoding unit that encodes each block based on the determined block type;
With
The encoding unit includes:
A block that specifies an encoding target block and peripheral blocks located around the encoding target block from the plurality of blocks, and specifies a position corresponding block corresponding to the peripheral block from the plurality of first encoding blocks. A specific part,
Omission to determine based on block information of the position-corresponding block whether or not to permit an abbreviated mode in which encoding of a part of information of the encoding target block is omitted as a block type of the encoding target block A mode permission section;
Only including,
The abbreviation mode permission unit
The image conversion characterized by permitting the abbreviated mode as a block type of the block to be encoded if the number of first encoded blocks included in the position corresponding block and intra-encoded is smaller than a predetermined number. apparatus. The image conversion device according to claim 7 ,
The encoding unit encodes uncompressed image data corresponding to an interlace in units of a block pair constituted by two blocks adjacent to each other in the vertical direction,
When the reference pair which is a block pair including the encoding target block is field-encoded, the block specifying unit is different from the reference pair and a block constituting a frame-encoded block pair is used as the peripheral block. An image conversion device characterized by specifying. The image conversion device according to claim 7 ,
The encoding unit encodes uncompressed image data corresponding to an interlace in units of a block pair constituted by two blocks adjacent to each other in the vertical direction,
The block specifying unit selects a first block pair that is different from the reference pair and is field-encoded when a reference pair that is a block pair including the encoding target block is field-encoded, and An image conversion apparatus that identifies a block whose field type matches the encoding target block among the two blocks constituting a block pair as the peripheral block. The image conversion apparatus according to claim 9 , wherein
The block specifying unit is
If the position-corresponding block is field-encoded and includes a first field and a second field, two first encodings corresponding to two blocks constituting the first block pair and including the position-corresponding block Identify the block,
The abbreviation mode permission unit
The image conversion apparatus according to claim 1, wherein if at least one of the two first encoded blocks is intra-encoded, the position corresponding block is determined to be intra-encoded. In the image conversion apparatus according to any one of claims 7 to 10,
The abbreviated mode is
An image conversion apparatus comprising: a skip mode in which encoding of motion information and pixel information of the encoding target block is omitted. In the image conversion apparatus according to any one of claims 7 to 10,
The abbreviated mode is
Direct mode in which encoding of motion information of the encoding target block is omitted,
An image conversion apparatus comprising: In the image conversion apparatus according to any one of claims 7 to 10,
The encoding target block is composed of a plurality of sub-blocks,
The abbreviated mode is
A mode in which encoding of motion information of each sub-block is omitted,
An image conversion apparatus comprising:

Images