JP3542572B2 - Image decoding method and apparatus - Google Patents

Description

量子化後、直流変換成分（以下、ＤＣ成分と記す）については、２０３で近傍サブブロック間で１次元予測され、２０４でその予測誤差をハフマン符号化する。ここでは、予測誤差の量子化出力をグループに分け、まず予測誤差の所属するグループの識別番号をハフマン符号化し、続いてグループ内のいずれかの値であるかを等長符号で表わす。
【００１４】
一方、ＤＣ成分以外の交流変換成分（以下、ＡＣ成分と記す）は、この量子化出力を、２０５で図８に示す様に低周波成分から高周波成分へとジグザグ走査しながら符号化する。即ち、量子化出力が“０”でない変換係数（以下、有意係数と記す）は、その値によりグループに分類され、２０６において、そのグループ識別番号と、直前の有意変換係数との間に挟まれた量子化出力が“０”の変換係数（以下、無効係数と記す）の個数とを組にしてハフマン符号化される。続いてグループ内のいずれの値であるかが等長符号で表わされる。
【００１５】
図１−１は本実施の形態の画像符号化装置の概略構成を示すブロック図である。
【００１６】
複数種の信号からなる画像信号Ｙ，Ｐｂ，Ｐｒのそれぞれが、端子１ａ，１ｂ，１ｃより入力される。ここでは複数の信号として、カラー画像信号の輝度信号Ｙと、色差信号Ｐｂ，Ｐｒの３つの信号が入力されるものとしている。各入力信号は、それぞれＡ／Ｄコンバータ３ａ，３ｂ，３ｃに入力されてデジタル信号に変換される。この実施の形態では、輝度信号Ｙは、色差信号Ｐｂ，Ｐｒの２倍のサンプリングレートでサンプリングされ、また図示しない線順次化回路で、色差信号は線順次化されるものとする。これにより、Ａ／Ｄ変換後の各信号のデータ量を比べると、輝度信号Ｙが“４”に対し、色差信号Ｐｂ、Ｐｒはそれぞれ“１”の割合となっている。
【００１７】
次に、このデジタル化された各データは、ブロック化回路５ａ，５ｂ，５ｃのそれぞれで、例えば８×８のブロック毎にまとめられる。各データは前述のように、４：１：１のデータ量比になっているため、生成される各ブロックの画面上での大きさは、輝度信号Ｙの１に対し、色差信号Ｐｂ，Ｐｒは４倍の大きさとなる。こうしてブロック化された各データは、従来技術で述べたような可変長圧縮方式により、符号化器７ａ，７ｂ，７ｃのそれぞれにおいて可変長圧縮符号化されるが、この実施の形態においては、特にここで各信号毎にそれぞれ画像データを複数の領域に分割し、これら各領域において閉じた可変長符号化を行なうようにしている。これらの領域としては、例えば前述の８×８のブロックを４０個集めたデータ量からなる領域とする。
【００１８】
上記のようにして、各信号毎に、又各領域毎に圧縮符号化された生成符号群はメモリ９に書込まれ、読み出し時に１つのデータ列にまとめられる。ここでは、アドレスコントローラ１１により、メモリ９よりのデータの読み出し順が制御されており、本実施の形態においては、前記複数の信号毎に、即ち、輝度信号Ｙ、色差信号Ｐｂ，Ｐｒ毎に、符号化された生成符号群が、画面上で同一の位置に属する領域毎、又は近傍位置に属する領域毎に時系列的にまとめて読み出される。この読み出し順については詳しく後述する。
【００１９】
メモリ９から読み出されたデータ列は、シンクコード付加部１３においてシンクコードが所定位置に挿入され、更に伝送ＩＤ付加部１５において伝送ＩＤが挿入される。１７は境界情報付加部であり、前述のように分割された領域に対する生成符号の区切りの情報をデータ列に、例えばマーカーコードのような形で挿入する。これより受信側でも各生成符号の区切りが検出でき、可変長符号の復号が、これら分割領域毎に確実に行なえるようになる。
【００２０】
更に、圧縮符号化された生成符号に対して、誤り検出・訂正符号化回路１９において誤り訂正符号化が行なわれ、誤り検出・訂正符号のパリティビットが前記データ列の所定位置に挿入されて伝送される。
【００２１】
２１は伝送路で、即時伝送であれば光ファイバ・衛星・マイクロ波等の地上電波・光空間等の伝送媒体であるし、蓄積伝送であれば、デジタルＶＴＲやＤＡＴ等のテープ上の媒体・フロッピーディスクや光ディスク等の円板状の媒体・半導体メモリ等の固体の媒体等の記憶媒体となる。この伝送レートについては、元の画像の情報量と圧縮率と要求する伝送時間とにより決定され、数十キロビット／秒から数十メガビット／秒まで様々である。
【００２２】
次に受信側の動作について図１−２を参照して説明する。
【００２３】
伝送路２１から受信したデータは、まずシンクコード検出部２３において同期検出され、伝送ＩＤ検出部２５において、ＩＤによりそのデータの属性が検出されて、その情報をもとにしたアドレスコントローラ２７の制御により、一旦メモリ２９に蓄えられる。メモリ２９のデータに対しては、誤り検出・訂正部３１において、データの誤り検出及び訂正が実施され、可能な限り伝送中に付加された誤りが取り除かれる。また、訂正しきれなかった誤りがある場合は、そのデータ群にフラグを立てておき、後段の補間回路３７ａ，３７ｂ，３７ｃにおいて補間処理が行なわれる。
【００２４】
そして、境界情報検出部３３により、分割された領域の圧縮符号部の境を検出し、その情報をもとにアドレスコントローラ２７により、メモリ２９からの読み出しアドレスを制御して、データを複数の信号、即ち本実施の形態においては輝度信号Ｙ，色差信号Ｐｂ，Ｐｒ毎に区分けし、かつ分割された領域毎に区分けしてメモリ２９から読み出す。こうして区分けされてメモリ２９より読み出されたデータは、復号器３５ａ，３５ｂ，３５ｃにおいてそれぞれ伸長・復号され、補間回路３７ａ，３７ｂ，３７ｃにおいて、訂正しきれなかった誤りが含まれるデータ群に対して分割領域単位で補間処理が実施される。補間処理の具体的方法としては、前フレームデータを用いた補間などがある。さらに、補間処理された後、データは逆ブロック回路３９ａ，３９ｂ，３９ｃで各信号毎に元の信号伝送順に戻され、色信号Ｐｂ，Ｐｒについては図示しない同時化回路で、線順次化されているデータが復元される。
【００２５】
そして各信号は、Ｄ／Ａ変換回路４１ａ，４１ｂ，４１ｃでアナログデータに変換されて、端子４３ａ，４３ｂ，４３ｃよりそれぞれ輝度信号Ｙ，色差信号Ｐｂ，Ｐｒの画像信号として出力される。ここで、復号器３５ａ，３５ｂ，３５ｃにおける可変長符号の伸長・復号に際しては、従来のように符号化された領域分割がなされていない場合には、一度誤りを起こすと、それ以降の復号処理が全く行なわれなくなってしまう。しかし、この実施の形態では前述したように、符号化時に画像データの領域を分割しており、その境界情報を付加して伝送しているために、復号処理よりの復帰を迅速に行なうことができる。
【００２６】
次に、図２、図３、図４、図５を参照して本実施の形態を更に詳しく説明する。
【００２７】
図２（Ａ）（Ｂ）は伝送対象の画像データの一例を示す図で、図２（Ａ）は１枚の画像の輝度信号Ｙを、横１２８０画素、縦１０８８画素、各画素を８ビットでＡ／Ｄ変換した画像とする。この場合における輝度信号Ｙの１枚当たりのデータ容量は、
１，２８０×１，０８８×８（ビット）＝ １１，１４１，１２０ビット
となる。
【００２８】
一方、前述のように色差信号Ｐｂ，Ｐｒのそれぞれは、輝度信号Ｙに対して１／２のサンプリングレートでサンプルされ、さらに色差線順次化しているため、１枚当たりのデータ容量は図２（Ｂ）に示すように、
６４０×５４４×８（ビット）＝ ２，７８５，２８０ビット
となる。
【００２９】
従って、輝度信号Ｙ，色差信号Ｐｂ，Ｐｒの合計では、１６，７１１，６８０（=11,141,120+2,785,280 ×2)ビットにより、１枚の画像が構成されていることになる。
【００３０】
さて、ここで（横８画素）×（縦８画素）をＤＣＴサブブロックとし、図２に示すように、伝送対象の１画像を各信号毎に、４０ＤＣＴサブブロックを１リシンクブロック（横３２０画素×縦８画素）として分割する。これを本実施の形態の分割領域として、各分割領域内で閉じた可変長符号化を行なう。尚、かかる分割領域は、本実施の形態に限らず他の分割方法であっても良い。
【００３１】
この場合、輝度信号Ｙに対しては、１画面分のデータはリシンクブロックにより、横４、縦１３６からなる合計５４４の領域に分割される。一方、色差信号Ｐｂ，Ｐｒに関しては、横２、縦６８の合計１３６の領域に各々分割される。ここで、１リシンクブロック当たりのデータ容量は、各信号とも、４０×８×８×８＝２０，４８０（ビット）であるが、１リシンクブロックの画面上でにおける大きさは、図２に示すように、（輝度信号リシンクブロック）：（色差信号ブロック）＝１：４となる。
【００３２】
図３は本実施の形態における伝送フォーマットの一例を示す図である。
【００３３】
前述したようにして符号化された符号化データは、リシンクブロック単位で区別することができるように、本実施の形態の境界情報（マーカーコード）が付加されるが、ここではリシンクブロックとリシンクブロックの境界にマーカーコードを挿入する場合の例を示している。なお、マーカーコードとしては、符号化データでは発生しえないビットパターンを割当てる必要がある。
【００３４】
このようにして境界情報が付加された符号化データは、更に誤り検出・訂正符号化されるが、ここでは１２８シンボル（以降、１シンンボル＝８ビットとする）のデータに対し、４シンボルのパリテイビットが付加されるものとする。このデータにシンクコード２シンボル、伝送ＩＤ２シンボルを付加したものが、伝送単位となっている。
【００３５】
ここで、符号化データは可変長符号化されているため、１リシンクブロック毎の符号化データ列の長さは一定ではなく、それぞれがまちまちの長さとなるため、１つのリシンクブロックのデータが、複数の誤り検出・訂正ブロックに跨がる場合もあり得る。逆に１つの誤り検出・訂正符号のブロックが、いくつものリシンクブロックのデータより構成される場合もあり、その数も一定とはならない。そして、１つの誤り検出・訂正符号のブロックが複数のリシンクブロックより構成される場合には、この誤り検出・訂正ブロック中の誤りが訂正不能となった時には、その影響は複数のリシンクブロックにまたがることとなり、当該複数のリシンクブロックが前述のような補間処理を施されることとなる。
【００３６】
本実施の形態においては、上記のように訂正不能の誤りが発生した場合においても、その影響が小さくなるように、符号化データの伝送順も規定している。
【００３７】
いま、図４−１及び図４−２のように各信号毎にリシンクブロックをナンバリングしたとすると、各信号毎のデータ量を考慮して、輝度信号Ｙの４リシンクブロックに対し、色差信号Ｐｂ，Ｐｒをそれぞれ１リシンクブロック伝送するため、例えば、
Ｙ（０，０），Ｙ（０，１），Ｐｂ（０，０），
Ｙ（０，２），Ｙ（０，３），Ｐｒ（０，０），
Ｙ（１，０），Ｙ（１，３），Ｐｂ（０，１），
Ｙ（１，２），Ｙ（１，３），Ｐｒ（０，１），……
と伝送したとすると、仮にＹ（０，２），Ｙ（０，３），Ｐｒ（０，０）が同一誤り検出・訂正ブロックに含まれ、その誤り検出・訂正ブロックが訂正不能となった場合には、画面上での補間領域が分散されてしまうこととなる。
【００３８】
そこで、本実施の形態においては、図５のように、
例えば、
Ｙ（０，０），Ｙ（０，１），Ｐｂ（０，０），
Ｙ（１，０），Ｙ（１，１），Ｐｒ（０，０），
Ｙ（０，２），Ｙ（０，３），Ｐｂ（０，１），
Ｙ（１，２），Ｙ（１，３），Ｐｒ（０，１），……
または、
Ｙ（０，０），Ｙ（０，１），Ｙ（１，０），
Ｙ（１，１），Ｐｂ（０，０），Ｐｒ（０，０），
Ｙ（０，２），Ｙ（０，３），Ｙ（１，２），
Ｙ（１，３），Ｐｒ（０，１），Ｐｒ（０，１），……
のように、画面上で同一位置もしくは、近傍位置にあるリシンクブロックをまとめて伝送するように、アドレスコントローラ１１を制御する。
【００３９】
これにより、当該誤り検出・訂正ブロックが訂正不能となっても、前述のように画面上での補間領域が分割される確率が少なくなる。
【００４０】
なお、本実施の形態においては、画像データを構成する複数種の信号は、輝度信号Ｙ，色差信号Ｐｂ，Ｐｒに限定されるものではなく、例えば、ＲＧＢ，ＹＵＶ，ＹＭＣＫ等の信号により構成されていてもよい。さらに、画像データを分割する領域の構成方法も、本実施の形態にあげた分割方法に限定されるものではない。
【００４１】
尚、本発明は、複数の機器から構成されるシステムに適用しても、１つの機器からなる装置に適用しても良い。又、本発明はシステム或は装置にプログラムを供給することによって達成される場合にも適用できることは言うまでもない。
【００４２】
以上説明したように本実施の形態によれば、圧縮効率において優れている可変長符号化方式の特徴を損なうことなく、伝送路に混入した誤りを、誤り検出・訂正符号で訂正しきれない場合においても、その影響を最小限に抑えることが可能となる。
【００４３】
即ち、伝送路上でデータに誤りが発生し、受信側でその誤りを訂正しきれない場合においても、その影響はリシンクブロック毎に収束させることができる。
【００４４】
更に、誤りを訂正しきれなかつた際に行なう補間処理については、補間処理が施されるリシンクブロックを画面上で位置的に分散させることなく、同一位置あるいは近傍位置にまとめることができ、画像データの劣化を最小限に抑えることが可能となり、人間の視覚上、劣化が気ならない、極めて良好な画像を再生できる画像符号化装置を提供することができる。
【００４５】
【発明の効果】
以上説明したように本発明によれば、所定数の輝度信号の大ブロックの可変長符号化データと、その符号化データと画像空間上で同一位置にある色差信号の大ブロックの可変長符号化データとを一単位にし、その単位が順次複数配列されており、かつ配列された大ブロック間に符号化データと識別可能な識別コードが挿入されている伝送ブロックを入力して復号し、その伝送ブロック内に誤りが発生した際、識別コードにより大ブロック間の境界を検出して前記符号化データを復号することにより、誤りを起こすとそれ以降の大ブロックの復号処理が全く行われなくなってしまうといった事態の発生を防止でき、画像データの劣化の領域を最小限に抑えて画像を再生できるという効果がある。
【図面の簡単な説明】
【図１−１】本実施の形態の画像符号化装置の概略構成を示すブロック図である。
【図１−２】本実施の形態の画像符号化装置の概略構成を示すブロック図である。
【図２】本実施の形態において符号化される画像データの構成を示す図である。
【図３】本実施の形態における伝送フォーマットの一例を示す図である。
【図４−１】本実施の形態における輝度信号Ｙのリシンクブロックの番号構成例を示す図である。
【図４−２】本実施の形態における色差信号のリシンクブロックの番号構成例を示す図である。
【図５】本実施の形態におけるリシンクブロックの伝送順序を示す図である。
【図６】本実施の形態の画像符号化装置の概略構成を示すブロック図である。
【図７】図６の可変長符号化方式を説明するための図である。
【図８】図６の可変長符号化方式を説明するための図である。
【符号の説明】
３ａ，３ｂ，３ｃ Ａ／Ｄコンバータ
５ａ，５ｂ，５ｃ ブロック化部
７ａ，７ｂ，７ｃ 符号化部
９，２９ メモリ
１１，２７ アドレスコントローラ
１３ シンクコード付加部
１５ 伝送ＩＤ付加部
１７ 境界情報付加部
１９ 誤り訂正符号化
２１ 伝送路
３５ａ，３５ｂ，３５ｃ 復号器
３７ａ，３７ｂ，３７ｃ 補間部
３９ａ，３９ｂ，３９ｃ 逆ブロック化部
４１ａ，４１ｂ，４１ｃ Ｄ／Ａコンバータ[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image decoding method and apparatus for inputting and decoding encoded image data.
[0002]
[Prior art]
Conventionally, when encoding image data, there are variable-length encoding and fixed-length encoding. In particular, in variable-length encoding, encoding such as predictive encoding is also known.
[0003]
[Problems to be solved by the invention]
However, such a variable-length coding scheme is excellent in compression efficiency, but if an error occurs in the compressed data on the transmission path, subsequent decoding cannot be performed at all. As a result, the image after the error has occurred in the compressed data may be disturbed, resulting in a very unsightly state.
[0004]
SUMMARY OF THE INVENTION The present invention has been made in view of the above-described conventional example, and provides an image decoding method and apparatus capable of performing efficient transmission and reproducing even if an error occurs in a transmission block while minimizing the area of image data deterioration. The purpose is to provide.
[0005]
[Means for Solving the Problems]
In order to achieve the above object, an image decoding device according to the present invention has the following configuration. That is,
Luminance signal and color difference signals constituting the image data is divided into small blocks of small blocks and a plurality of color difference signals of the respective plurality of luminance signals are variable length encoded, large luminance signal small blocks of a plurality of said luminance signal and block, the small block of the plurality of the color difference signals as a large block of color difference signals, 1 and large block of coded data of the luminance signal of a predetermined number less than the screen, the large block of the predetermined number of the luminance signal And a predetermined number of coded data of the large blocks of the color difference signal at the same position on the image space as one unit, the plurality of units are sequentially arranged, and the code is arranged between the arranged large blocks. An image decoding apparatus for inputting and decoding a transmission block in which an identification code that can be identified as encoded data is inserted,
Separating means for separating the encoded data of the luminance signal and the encoded data of the color difference signal from the input transmission block,
Decoding means for decoding the encoded data of the luminance signal and the encoded data of the color difference signal separated by the separating means,
When an error occurs in the transmission block, the decoding unit detects a boundary between large blocks based on the identification code and decodes the encoded data .
[0006]
In order to achieve the above object, an image decoding method according to the present invention includes the following steps. That is,
Luminance signal and color difference signals constituting the image data is divided into small blocks of small blocks and a plurality of color difference signals of the respective plurality of luminance signals are variable length encoded, large luminance signal small blocks of a plurality of said luminance signal and block, the small block of the plurality of the color difference signals as a large block of color difference signals, 1 and large block of coded data of the luminance signal of a predetermined number less than the screen, the large block of the predetermined number of the luminance signal And a predetermined number of coded data of the large blocks of the color difference signal at the same position on the image space as one unit, the plurality of units are sequentially arranged, and the code is arranged between the arranged large blocks. An image decoding method for inputting and decoding a transmission block in which an identification code that can be identified as encoded data is inserted,
A separation step of separating encoded data of the luminance signal and encoded data of the color difference signal from the input transmission block,
Decoding the encoded data of the luminance signal and the encoded data of the color difference signal separated in the separation step,
In the decoding step, when an error occurs in the transmission block, a boundary between large blocks is detected based on the identification code, and the encoded data is decoded.
[0007]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.
[0008]
FIG. 6 is a block diagram illustrating a schematic configuration of the image encoding device according to the present embodiment.
[0009]
In FIG. 6, image data input from a terminal 101 is converted from an analog signal to a digital signal in an A / D converter 102. The converted digital signal is compression-encoded in a variable length by the encoding unit 103. Then, the encoded code is transmitted to the transmission path 105 by the error correction encoding unit 104 added with a parity code for later error correction. The data received from the transmission path 105 is temporarily stored in a memory unit 106, and an error is corrected by an error correction unit 107. The decoding unit 108 expands and decodes the variable-length data read from the memory 106. After the decoded signal is converted from a digital signal to an analog signal by the D / A converter 109, It is output as an image signal.
[0010]
As a color image compression method of the encoding unit 103 in FIG. 6, various methods have been proposed, and a typical color image encoding method is a method called ADCT method. This ADCT method is described in “The Still Image Coding Method” by Takahiro Saito et al. In the Journal of the Institute of Television Engineers of Japan (Vol.44, No.2 (1990)), and Tomohiro Ochi in Preprints 14 These are described in detail in “International Standard Trends in Still Image Coding” and the like.
[0011]
FIG. 7 is a diagram illustrating a configuration concept of the encoding unit 103 using the ADCT method.
[0012]
The input image is 8-bit data output from the A / D converter 102 shown in FIG. 6, that is, data converted into 256 gradations / colors. The number of colors is three colors such as RGB, YUV, YPbPr, and YMCK. Alternatively, four colors are used. Here, the input image is immediately subjected to two-dimensional discrete cosine transform (hereinafter referred to as DCT) in units of 8 × 8 pixel sub-blocks, and then the linear quantization unit 202 performs linear quantization of the transform coefficients. At the time of this quantization, the quantization step size is different for each transform coefficient, and the quantization step size for each transform coefficient is 8 × 8 quantization taking into account the difference in visibility for quantization noise for each transform coefficient. the matrix elements and 2 ^S multiplied value. Here, S is 0 or a positive or negative integer, and is called a scaling factor. The image quality and the amount of generated data are controlled by the value of S. Table 1 shows an example of the quantization matrix element.
[0013]
[Table 1]

After quantization, a DC conversion component (hereinafter, referred to as a DC component) is one-dimensionally predicted between neighboring sub-blocks at 203, and the prediction error is Huffman-coded at 204. Here, the quantized output of the prediction error is divided into groups, first, the identification number of the group to which the prediction error belongs is Huffman-coded, and subsequently, any value in the group is represented by an equal-length code.
[0014]
On the other hand, an AC conversion component other than the DC component (hereinafter, referred to as an AC component) encodes the quantized output while performing zigzag scanning from a low-frequency component to a high-frequency component at 205 as shown in FIG. That is, a transform coefficient whose quantized output is not “0” (hereinafter, referred to as a significant coefficient) is classified into a group by its value, and is interposed at 206 between the group identification number and the immediately preceding significant transform coefficient. The Huffman coding is performed by combining the number of transform coefficients (hereinafter, referred to as invalid coefficients) whose quantized output is “0”. Subsequently, which value in the group is represented by an isometric code.
[0015]
FIG. 1-1 is a block diagram showing a schematic configuration of the image coding apparatus according to the present embodiment.
[0016]
Image signals Y, Pb, and Pr composed of a plurality of types of signals are input from terminals 1a, 1b, and 1c. Here, it is assumed that three signals of a luminance signal Y of a color image signal and color difference signals Pb and Pr are input as a plurality of signals. Each input signal is input to each of the A / D converters 3a, 3b, and 3c, and is converted into a digital signal. In this embodiment, the luminance signal Y is sampled at twice the sampling rate of the color difference signals Pb and Pr, and the color difference signals are line-sequentialized by a line-sequencing circuit (not shown). As a result, when the data amount of each signal after the A / D conversion is compared, the ratio of the luminance signal Y is “4” and the color difference signals Pb and Pr are each “1”.
[0017]
Next, each of the digitized data is put together in each of the blocking circuits 5a, 5b, 5c, for example, for each 8 × 8 block. Since each data has a data amount ratio of 4: 1: 1 as described above, the size of each generated block on the screen is such that the luminance signal Y is 1 and the color difference signals Pb, Pr Is four times larger. Each of the data thus blocked is subjected to variable-length compression encoding in each of the encoders 7a, 7b, and 7c according to the variable-length compression method as described in the related art. Here, the image data is divided into a plurality of regions for each signal, and closed variable length coding is performed in each of these regions. These regions are, for example, regions having a data amount of 40 forty-eight 8 × 8 blocks.
[0018]
As described above, the generated code group that has been compression-coded for each signal and for each region is written to the memory 9 and is combined into one data string at the time of reading. Here, the order of reading data from the memory 9 is controlled by the address controller 11, and in the present embodiment, for each of the plurality of signals, that is, for each of the luminance signal Y and the color difference signals Pb and Pr. The encoded generated code group is read out collectively in time series for each region belonging to the same position on the screen or for each region belonging to a nearby position. This reading order will be described later in detail.
[0019]
In the data sequence read from the memory 9, the sync code is inserted at a predetermined position in the sync code adding unit 13, and the transmission ID is inserted in the transmission ID adding unit 15. Reference numeral 17 denotes a boundary information adding unit that inserts information of a generated code delimiter for the divided area into the data string in the form of, for example, a marker code. As a result, the receiving side can detect the break of each generated code, and can reliably decode the variable length code for each of the divided areas.
[0020]
Further, the generated code that has been compression-coded is subjected to error correction coding in an error detection / correction coding circuit 19, and a parity bit of the error detection / correction code is inserted into a predetermined position of the data string and transmitted. Is done.
[0021]
Reference numeral 21 denotes a transmission line, which is a transmission medium such as a terrestrial radio wave such as an optical fiber, a satellite, or a microwave for immediate transmission, and a transmission medium such as a digital VTR or DAT for storage transmission. It is a storage medium such as a disk-shaped medium such as a floppy disk or an optical disk, or a solid medium such as a semiconductor memory. The transmission rate is determined by the information amount of the original image, the compression rate, and the required transmission time, and varies from several tens of kilobits / second to several tens of megabits / second.
[0022]
Next, the operation on the receiving side will be described with reference to FIG.
[0023]
The data received from the transmission line 21 is first synchronously detected by the sync code detection unit 23, and the transmission ID detection unit 25 detects the attribute of the data by the ID, and controls the address controller 27 based on the information. Is temporarily stored in the memory 29. The data in the memory 29 is subjected to data error detection and correction in the error detection / correction unit 31, and errors added during transmission are removed as much as possible. If there is an error that cannot be corrected, a flag is set in the data group, and interpolation processing is performed in the interpolation circuits 37a, 37b, and 37c at the subsequent stage.
[0024]
Then, the boundary information detection unit 33 detects the boundary of the compression code part of the divided area, and based on the information, controls the read address from the memory 29 by the address controller 27 to convert the data into a plurality of signals. That is, in the present embodiment, the data is divided from the luminance signal Y and the color difference signals Pb and Pr, and divided from the divided areas and read out from the memory 29. The data thus divided and read out from the memory 29 are decompressed and decoded by the decoders 35a, 35b, and 35c, respectively, and are interpolated by the interpolation circuits 37a, 37b, and 37c. The interpolation process is performed for each divided area. As a specific method of the interpolation processing, there is interpolation using the previous frame data and the like. Further, after the interpolation processing, the data is returned to the original signal transmission order for each signal by the inverse block circuits 39a, 39b, 39c, and the color signals Pb, Pr are line-sequentially converted by a synchronization circuit (not shown). Data is restored.
[0025]
Each signal is converted into analog data by D / A conversion circuits 41a, 41b, and 41c, and output as luminance signal Y and color difference signals Pb and Pr from terminals 43a, 43b, and 43c, respectively. Here, when the variable-length codes are decompressed and decoded by the decoders 35a, 35b, and 35c, if an encoded region is not divided as in the related art, if an error occurs once, subsequent decoding processing is performed. Will not be performed at all. However, in this embodiment, as described above, since the image data area is divided at the time of encoding and the boundary information is added and transmitted, it is possible to quickly return from the decoding processing. it can.
[0026]
Next, the present embodiment will be described in more detail with reference to FIGS. 2, 3, 4, and 5. FIG.
[0027]
2A and 2B are diagrams showing an example of image data to be transmitted. FIG. 2A shows a luminance signal Y of one image as 1280 horizontal pixels, 1088 vertical pixels, and each pixel is 8 bits. The image is A / D converted. In this case, the data capacity per luminance signal Y is:
1,280 × 1,088 × 8 (bits) = 11,141,120 bits.
[0028]
On the other hand, as described above, each of the color difference signals Pb and Pr is sampled at a sampling rate of に 対 し て with respect to the luminance signal Y, and is further subjected to color difference line sequential processing. As shown in B),
640 × 544 × 8 (bits) = 2,785,280 bits.
[0029]
Therefore, in the sum of the luminance signal Y and the color difference signals Pb and Pr, one image is composed of 16,711,680 (= 11,141,120 + 2,785,280 × 2) bits.
[0030]
Now, here, (8 horizontal pixels) × (8 vertical pixels) are DCT sub-blocks. As shown in FIG. 2, one image to be transmitted is divided into 40 DCT sub-blocks by one resync block (320 horizontal pixels) for each signal. X 8 pixels). This is set as a divided region in the present embodiment, and variable length coding closed in each divided region is performed. Note that such a divided area is not limited to the present embodiment, and may be another divided method.
[0031]
In this case, for the luminance signal Y, the data for one screen is divided by the resync block into a total of 544 areas of 4 horizontal and 136 vertical. On the other hand, the color difference signals Pb and Pr are each divided into a total of 136 areas of horizontal 2 and vertical 68. Here, the data capacity per resync block is 40 × 8 × 8 × 8 = 20,480 (bits) for each signal, but the size of one resync block on the screen is shown in FIG. Thus, (luminance signal resync block) :( color difference signal block) = 1: 4.
[0032]
FIG. 3 is a diagram illustrating an example of a transmission format according to the present embodiment.
[0033]
The coded data coded as described above is added with the boundary information (marker code) according to the present embodiment so that it can be distinguished in units of resync blocks. Shows an example in which a marker code is inserted at the boundary of. It is necessary to assign a bit pattern that cannot occur in the encoded data as the marker code.
[0034]
The encoded data to which the boundary information is added in this manner is further subjected to error detection / correction encoding. In this case, 128 symbols (hereinafter, 1 symbol = 8 bits) of data are converted to 4 symbols of parity data. It is assumed that a tail bit is added. The transmission unit is obtained by adding two sync code symbols and two transmission ID symbols to this data.
[0035]
Here, since the coded data is variable-length coded, the length of the coded data sequence for each resync block is not constant, and each data length varies, so that the data of one resync block is There may be a case where a plurality of error detection / correction blocks are straddled. Conversely, one error detection / correction code block may be composed of data of several resync blocks, and the number thereof is not constant. When one error detection / correction code block is composed of a plurality of resync blocks, when an error in the error detection / correction block becomes uncorrectable, the effect extends to a plurality of resync blocks. That is, the plurality of resync blocks are subjected to the interpolation processing as described above.
[0036]
In the present embodiment, even if an uncorrectable error occurs as described above, the transmission order of the encoded data is defined so that the effect is reduced.
[0037]
Now, assuming that resync blocks are numbered for each signal as shown in FIGS. 4-1 and 4-2, the color difference signal Pb is applied to the four resync blocks of the luminance signal Y in consideration of the data amount of each signal. , Pr are transmitted by one resync block, for example,
Y (0,0), Y (0,1), Pb (0,0),
Y (0,2), Y (0,3), Pr (0,0),
Y (1,0), Y (1,3), Pb (0,1),
Y (1,2), Y (1,3), Pr (0,1), ...
Suppose that Y (0,2), Y (0,3), Pr (0,0) are included in the same error detection / correction block, and the error detection / correction block cannot be corrected. In such a case, the interpolation areas on the screen will be dispersed.
[0038]
Therefore, in the present embodiment, as shown in FIG.
For example,
Y (0,0), Y (0,1), Pb (0,0),
Y (1,0), Y (1,1), Pr (0,0),
Y (0,2), Y (0,3), Pb (0,1),
Y (1,2), Y (1,3), Pr (0,1), ...
Or
Y (0,0), Y (0,1), Y (1,0),
Y (1,1), Pb (0,0), Pr (0,0),
Y (0,2), Y (0,3), Y (1,2),
Y (1,3), Pr (0,1), Pr (0,1), ...
As described above, the address controller 11 is controlled so that resync blocks located at the same position or near positions on the screen are transmitted collectively.
[0039]
As a result, even if the error detection / correction block becomes uncorrectable, the probability that the interpolation area on the screen is divided as described above is reduced.
[0040]
In the present embodiment, the plurality of types of signals constituting the image data are not limited to the luminance signal Y and the color difference signals Pb and Pr, but are composed of signals such as RGB, YUV, and YMCK. May be. Further, the method of configuring the area into which the image data is divided is not limited to the dividing method described in the present embodiment.
[0041]
The present invention may be applied to a system including a plurality of devices or to an apparatus including a single device. Needless to say, the present invention can be applied to a case where the present invention is achieved by supplying a program to a system or an apparatus.
[0042]
As described above, according to the present embodiment, the error mixed in the transmission path cannot be corrected by the error detection / correction code without impairing the characteristics of the variable-length coding scheme having excellent compression efficiency. In this case, the influence can be minimized.
[0043]
That is, even if an error occurs in the data on the transmission path and the error cannot be corrected on the receiving side, the effect can be converged for each resync block.
[0044]
Furthermore, regarding the interpolation processing to be performed when an error cannot be corrected, the resync blocks to be subjected to the interpolation processing can be grouped at the same position or a nearby position without dispersing the positions on the screen. This makes it possible to provide an image encoding apparatus that can reproduce extremely good images without any noticeable deterioration in human vision.
[0045]
【The invention's effect】
As described above, according to the present invention, variable-length encoded data of a large block of a predetermined number of luminance signals and variable-length encoded data of a large block of a color difference signal at the same position in the image space as the encoded data Data is made into one unit, a plurality of units are sequentially arranged, and a transmission block in which an identification code identifiable with coded data is inserted between the arranged large blocks is input and decoded, and the transmission is performed. When an error occurs in a block, the boundary between large blocks is detected by the identification code and the encoded data is decoded. If an error occurs, decoding processing of the subsequent large block is not performed at all. Such a situation can be prevented , and there is an effect that an image can be reproduced while minimizing the area of image data deterioration.
[Brief description of the drawings]
FIG. 1-1 is a block diagram illustrating a schematic configuration of an image encoding device according to the present embodiment.
FIG. 1-2 is a block diagram illustrating a schematic configuration of an image encoding device according to the present embodiment.
FIG. 2 is a diagram showing a configuration of image data to be encoded in the present embodiment.
FIG. 3 is a diagram illustrating an example of a transmission format according to the present embodiment.
FIG. 4-1 is a diagram illustrating an example of a number configuration of a resync block of a luminance signal Y according to the present embodiment.
FIG. 4-2 is a diagram illustrating an example of a number configuration of a resync block of a color difference signal according to the present embodiment.
FIG. 5 is a diagram showing a transmission order of resync blocks in the present embodiment.
FIG. 6 is a block diagram illustrating a schematic configuration of an image encoding device according to the present embodiment.
FIG. 7 is a diagram for explaining the variable-length coding system of FIG.
FIG. 8 is a diagram for explaining the variable-length coding system of FIG.
[Explanation of symbols]
3a, 3b, 3c A / D converters 5a, 5b, 5c Blocking units 7a, 7b, 7c Encoding unit 9, 29 Memory 11, 27 Address controller 13 Sync code adding unit 15 Transmission ID adding unit 17 Boundary information adding unit 19 Error Correction Coding 21 Transmission Lines 35a, 35b, 35c Decoders 37a, 37b, 37c Interpolators 39a, 39b, 39c Deblockers 41a, 41b, 41c D / A Converters

Claims (2)

Luminance signal and color difference signals constituting the image data is divided into small blocks of small blocks and a plurality of color difference signals of the respective plurality of luminance signals are variable length encoded, large luminance signal small blocks of a plurality of said luminance signal and block, the small block of the plurality of the color difference signals as a large block of color difference signals, 1 and large block of coded data of the luminance signal of a predetermined number less than the screen, the large block of the predetermined number of the luminance signal And a predetermined number of coded data of the large blocks of the color difference signal at the same position on the image space as one unit, the plurality of units are sequentially arranged, and the code is arranged between the arranged large blocks. An image decoding apparatus for inputting and decoding a transmission block in which an identification code that can be identified as encoded data is inserted,
Separating means for separating the encoded data of the luminance signal and the encoded data of the color difference signal from the input transmission block,
Decoding means for decoding the encoded data of the luminance signal and the encoded data of the color difference signal separated by the separating means,
An image decoding apparatus according to claim 1, wherein when an error occurs in the transmission block, the decoding unit detects a boundary between large blocks based on the identification code and decodes the encoded data . Luminance signal and color difference signals constituting the image data is divided into small blocks of small blocks and a plurality of color difference signals of the respective plurality of luminance signals are variable length encoded, large luminance signal small blocks of a plurality of said luminance signal and block, the small block of the plurality of the color difference signals as a large block of color difference signals, 1 and large block of coded data of the luminance signal of a predetermined number less than the screen, the large block of the predetermined number of the luminance signal And a predetermined number of coded data of the large blocks of the color difference signal at the same position on the image space as one unit, the plurality of units are sequentially arranged, and the code is arranged between the arranged large blocks. An image decoding method for inputting and decoding a transmission block in which an identification code that can be identified as encoded data is inserted,
A separation step of separating encoded data of the luminance signal and encoded data of the color difference signal from the input transmission block,
Decoding the encoded data of the luminance signal and the encoded data of the color difference signal separated in the separation step,
The image decoding method according to claim 1, wherein, when an error occurs in the transmission block, a boundary between large blocks is detected by the identification code to decode the encoded data .

Images