WO1998019461A1 - Image encoder, image encoding method, image decoder, image decoding method, and data recording medium - Google Patents

Image encoder, image encoding method, image decoder, image decoding method, and data recording medium Download PDF

Info

Publication number
WO1998019461A1
WO1998019461A1 PCT/JP1997/003917 JP9703917W WO9819461A1 WO 1998019461 A1 WO1998019461 A1 WO 1998019461A1 JP 9703917 W JP9703917 W JP 9703917W WO 9819461 A1 WO9819461 A1 WO 9819461A1
Authority
WO
WIPO (PCT)
Prior art keywords
pixel
image
pixel value
value
encoded
Prior art date
Application number
PCT/JP1997/003917
Other languages
English (en)
French (fr)
Inventor
Shinya Kadono
Original Assignee
Matsushita Electric Industrial Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Matsushita Electric Industrial Co., Ltd. filed Critical Matsushita Electric Industrial Co., Ltd.
Priority to US09/077,378 priority Critical patent/US6173078B1/en
Priority to JP52029098A priority patent/JP3176378B2/ja
Priority to EP97909663A priority patent/EP0873015B1/en
Priority to DE69709847T priority patent/DE69709847T2/de
Publication of WO1998019461A1 publication Critical patent/WO1998019461A1/ja
Priority to US09/867,610 priority patent/US6487315B2/en
Priority to US10/238,600 priority patent/US6748117B2/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/593Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/41Bandwidth or redundancy reduction

Definitions

  • Image coding apparatus and image coding method Description Image coding apparatus and image coding method, image decoding apparatus and image decoding method, and data recording medium
  • the present invention relates to an image encoding device and an image encoding method, an image decoding device and an image decoding method, and a data recording medium, and in particular, to recording or transmitting an image signal with a smaller number of bits to impair image quality.
  • the present invention relates to an encoding process and a decoding process for performing the encoding process and a recording medium storing a program for realizing the encoding process or the decoding process.
  • Encoding processing in units of blocks is a method in which one image display area is divided into a plurality of blocks, and encoding processing of an input image signal (hereinafter, also referred to as an image input signal) is performed for each block.
  • the one image display area corresponds to one display screen in the encoding process based on MPEG2, and in the encoding process based on MPEG4, the This corresponds to a display area having a shape and a size corresponding to an object (object).
  • Each block is a display area including a predetermined number of pixels in one image display area, and a rectangular shape that is easy to process is often used as the shape of the block.
  • the encoding process of the image input signal corresponding to one image display area is completed for each block. Therefore, the image input signal is subjected to an encoding process, Even if a transmission error occurs when transmitting the obtained image coded signal, there is an advantage that the effect of the error can be converged in block units.
  • the block-based coding method has the following disadvantages.
  • the pixel correlation between blocks that is, the correlation between the pixel values existing between different blocks. It is difficult to use when.
  • a pixel value of a pixel to be coded (a pixel value to be coded) is converted to a pixel value of a plurality of coded pixels coded earlier. (Encoded pixel value), and performs a process of adaptively encoding the coded pixel value using the predicted pixel value.
  • the encoded pixel value referred to when encoding the encoded pixel value is limited to the pixels in the block.
  • the number of coded pixel values referenced is reduced. For this reason, the likelihood of the prediction value of the pixel to be coded becomes small, and the coding efficiency does not become too high.
  • the pixel-by-pixel coding method is a method of coding an input image signal for each pixel.
  • the coding process of the image input signal can be changed in pixel units. . Therefore, in this encoding method, for example, if universal encoding processing such as adaptive arithmetic encoding in which a code word is automatically updated for each pixel in accordance with the characteristics of an image input signal is performed, Such special image signals can be encoded with very high encoding efficiency.
  • the image-encoded signal obtained by the pixel-based encoding method for performing the universal encoding process described above includes a decoding process for updating the codeword on the decoding side in exactly the same way as the encoding side. Therefore, if a transmission error occurs when transmitting the above-described coded image signal, the decoding side cannot correctly perform decoding processing of the coded image signal due to the influence of the transmission error. Has the disadvantage that it lasts for a long time.
  • this encoding method is called a combination encoding method.
  • Fig. 13 (a) shows a state in which one frame screen is divided into a plurality of rectangular blocks
  • Fig. 13 (b) shows the inside of a block, in particular, the encoding target.
  • 3 shows an arrangement of pixels in a block and its peripheral blocks. Needless to say, the pixels are arranged in a matrix along each horizontal scanning line in one frame screen.
  • FG is a screen corresponding to one frame
  • B1 is a coded block that has already been subjected to coding processing on an image signal
  • BX is a block to be coded that is to be coded
  • B0 is This is an uncoded block on which no coding process has been performed on the image signal.
  • BLu, BLs, BLh, and 8111 are the upper, lower, left, and right boundaries of the block to be coded on one frame screen. Solid circles indicate coded pixels for which the pixel value has been coded, and dotted circles indicate uncoded pixels for which the pixel value has not been coded. Is shown.
  • Each of the blocks B is an image display area composed of 4 ⁇ 4 pixels on the one-frame screen FG.
  • FIG. 14 shows the positional relationship between the pixel to be coded PX to be coded and the peripheral pixels P0 to P9 located therearound. Are referred to when predicting the pixel value of the pixel to be encoded PX, and are hereinafter referred to as reference pixels P0 to P9.
  • the reference pixels P 8 and P 9 are pixels located on the same horizontal scanning line as the above-mentioned encoded pixel PX, and the reference pixels P 9 and P 8 are respectively the above-mentioned encoded pixels P 8 and P 9. It is located one or two pixels before x.
  • the positions of the reference pixels P5 and P1 in the horizontal direction on the one-frame screen FG are the same as the above-mentioned pixel to be encoded PX, and the reference pixels P5 and P1 are respectively It is located on the horizontal scanning line one and two pixels above the above-mentioned pixel to be encoded PX.
  • the reference pixels P 3, P 4, P 6, and P 7 are pixels located on the same horizontal scanning line as the reference pixel P 5, and the reference pixels P 4 and P 3 respectively correspond to the above-described encoded pixels.
  • the reference pixels P6 and P7 are located one and two pixels behind the coded pixel PX, respectively, one pixel and two pixels before the coded pixel PX.
  • the reference pixels P 0 and P 2 are pixels located on the same horizontal scanning line as the reference pixel P 1, and the reference pixel P 0 is located one pixel before the reference pixel P 1.
  • the reference pixel P2 is located one pixel behind the reference pixel P1.
  • an image signal corresponding to a one-frame screen FG is generated by a plurality of blocks B constituting the one-frame screen. And the coding process of the divided image signal is performed for each block.
  • the block-based encoding process includes a horizontal process of sequentially encoding the pixel values of each pixel from left to right along a horizontal pixel column in block B, The process is completed by sequentially performing the horizontal pixel row from the top row to the bottom row.
  • the pixel value of the pixel to be encoded Px is adapted from the pixel values of the reference pixels P0 to P9 located around the pixel to be encoded. Codewords to be used in the coding process of the pixel to be coded, according to the predicted value obtained by the prediction.
  • FIG. 16 (a), FIG. 16 (b) and FIG. 17 are diagrams for explaining the combination decoding method corresponding to the combination coding method described above.
  • B x ' is the decrypted block
  • B 1' is the decrypted block
  • B 0 ' is the undecoded block
  • B Lu', BL s', B Lh ', B Lm' are above, below, left, and right of the block to be decoded
  • B x ' ⁇ ⁇ ⁇ ⁇ ⁇ ′ to ⁇ 9 ′ are reference pixels corresponding to the decoded pixel ⁇ ⁇ ′.
  • an image signal corresponding to one frame screen FG ′ forms the one frame screen.
  • the image signal is divided so as to correspond to each of the plurality of blocks B ', and decoding processing of the divided image signal is performed for each block.
  • the block-based decryption process includes a horizontal process of sequentially decoding the pixel values of each pixel from left to right along a horizontal pixel column in block B ′. The process is completed by performing the processing from the top row to the bottom row sequentially for all the horizontal pixel rows.
  • the pixel value of the pixel to be decoded ⁇ ⁇ ′ is set to the value of each of the reference pixels ⁇ 0 ′ to ⁇ 9 ′ located around the pixel to be decoded. It adaptively predicts from the pixel value, and adaptively selects a codeword to be used for decoding the pixel to be decoded ⁇ ⁇ ′ according to the predicted value obtained by the prediction.
  • the combination encoding method that combines both the encoding processing in block units and the encoding processing in pixel units has the following problems.
  • the decoding method used in the decoding process is used.
  • the predicted value of the pixel ⁇ ⁇ ′ is used in the encoding process. It is necessary to match the predicted value of the pixel to be encoded ⁇ corresponding to the coded pixel P.
  • the reference pixel value referred to when generating the predicted value of the pixel to be encoded ⁇ X is equal to the pixel to be decoded ⁇ ⁇ ⁇ ′ corresponding to the pixel to be encoded ⁇ X. That is, it is necessary to completely match the reference pixel value that is referred to when generating the predicted value.
  • the decoding side when performing the encoding process on the pixel to be encoded ⁇ ⁇ , the unencoded pixels among the reference pixels ⁇ 0 to ⁇ 9 with respect to the pixel to be encoded ⁇ X
  • the decoding side performs the decoding on the decoding side as shown in FIG.
  • the prediction value of the decrypted pixel ⁇ ⁇ ' is determined by referring to the pixel values of the reference pixels ⁇ ⁇ 'to ⁇ 9' corresponding to the pixel ⁇ ⁇ '
  • pixel values are not obtained for the undecoded pixels ⁇ 2 ′, P 6 ′ and ⁇ 1 ′ among the reference pixels ⁇ ⁇ ⁇ to ⁇ 9 ′.
  • the pixel value of the decoded pixel ⁇ X ′ corresponding to the encoded pixel ⁇ X cannot be decoded.
  • the decoding process becomes difficult when the reference pixels ⁇ 0 to ⁇ 9 corresponding to the coded pixel ⁇ X include the uncoded pixels as described above.
  • the pixel value is regarded as a preset fixed value (for example, 0), and a predicted value of the coded pixel ⁇ ⁇ is generated.
  • a countermeasure has been taken to perform encoding processing on the pixel to be encoded ⁇ X.
  • the decoding side can correctly perform the decoding process using the predicted values for all the pixels in the block. Since the pixel value of the reference pixel, which is an uncoded pixel, is uniformly replaced with a fixed value, the correlation of the pixel value between the uncoded pixel and the coded pixel is lost, and as a result, the A problem arises in that the prediction efficiency of the coded pixel, that is, the certainty of the predicted value for the coded pixel is deteriorated.
  • the present invention has been made in order to solve the above-described problems, and performs adaptive encoding processing in units of pixels and encoding processing in units of blocks as uncoded pixels.
  • An image processing apparatus and an image processing method which can be combined without impairing the correlation of pixel values between the pixels to be encoded and while avoiding difficulty in decoding an image encoded signal;
  • An object of the present invention is to obtain a data storage medium storing an image processing program for realizing an image processing method.
  • An image processing apparatus is an image processing apparatus which sequentially encodes pixel values constituting an image signal based on pixel values of a plurality of peripheral pixels located around a pixel to be encoded.
  • An encoding device wherein an image signal composed of a plurality of pixel values corresponding to one image display area is blocked so as to correspond to a block composed of a predetermined number of pixels, and a predetermined number of pixels constituting an image signal of each block.
  • a block generating means for outputting a pixel value in block units, and a pixel value of an uncoded pixel among the peripheral pixels based on a pixel value of an encoded pixel among the peripheral pixels based on a predetermined rule.
  • Pixel value replacement means for replacing the obtained pseudo pixel value with an image signal comprising a plurality of pixel values corresponding to each of the blocks; and receiving each pixel value with the pixel value of the coded pixel and the pixel value of the uncoded pixel. Based on the pseudo pixel value
  • the encoding process for sequentially encoding had to go to each proc is obtained example Bei and encoding means for outputting the coded image signal.
  • the pixel value of the uncoded pixel among the plurality of peripheral pixels corresponding to the pixel to be coded is calculated as the pixel value of the coded pixel among the plurality of peripheral pixels. Since pixel value replacement means for replacing the pixel value with a pseudo pixel value obtained based on the value is provided, the process of coding the image signal corresponding to one image display area in units of blocks When the value is referred to pixel by pixel, even if the pixel to be coded is located adjacent to the boundary of the block and the reference pixels include uncoded pixels, Pixel image As the prime value, a pseudo pixel value correlated with the pixel values of other peripheral pixels can be referred to.
  • the decoding side also refers to the undecoded data that is referred to when decoding the coded pixel.
  • a pseudo pixel value obtained from the pixel value of the decoded pixel can be referenced instead of the pixel value.
  • the adaptive coding processing in pixel units and the coding processing in block units can be performed without impairing the correlation of the pixel values between the uncoded pixels and the to-be-coded pixels, and moreover, in a multiple manner. Combinations can be made while avoiding the difficulty of decoding encoded signals on the encoding side.
  • the pixel value replacement unit is configured to include a coded pixel having a shortest spatial distance from the uncoded pixel.
  • the pixel value of the pixel is set as a pseudo pixel value of the uncoded pixel.
  • the pixel value of the coded pixel closest to the uncoded pixel is used as the pseudo pixel value of the uncoded pixel.
  • the pixel value can have a strong correlation with the pixel values of the peripheral pixels.
  • the pixel value replacement unit is configured so that a spatial distance from the uncoded pixel is the shortest.
  • the pixel value of an encoded pixel located on the same horizontal scanning line as the encoded pixel is set as a pseudo pixel value of the uncoded pixel.
  • the encoding unit includes a pixel value of the encoded pixel and a pseudo pixel value of the uncoded pixel.
  • a predicted value generator for generating a predicted pixel value for the pixel to be encoded based on the following equation; and a difference value between the pixel value of the pixel to be encoded and the predicted pixel value of the pixel to be encoded.
  • an encoder that outputs the difference value as an image coded signal in block units.
  • the difference value between the pixel value of the above-mentioned encoded pixel and its predicted pixel value is encoded, so that it is necessary to encode the pixel value of the encoded pixel.
  • the coding amount can be reduced and the coding efficiency can be improved.
  • the present invention (Claim 5) is the image encoding device according to Claim 1 or 4, wherein the encoding means comprises: a pixel value of the encoded pixel; The code word for encoding the pixel value of the pixel to be encoded is selected based on the pseudo pixel value of the pixel.
  • a codeword for coding the pixel value of the pixel to be coded based on the pixel value of the coded pixel and the pseudo pixel value of the uncoded pixel Is selected so that in the encoding process of the image signal, the codeword is adaptively changed for each pixel, so that efficient encoding process can be realized.
  • the present invention is the image encoding device according to Claim 1 or 4, wherein the encoding means comprises: a pixel value of the encoded pixel; Based on the pseudo pixel value of the coded pixel, a probability table corresponding to a code code for arithmetically coding the pixel value of the coded pixel is selected, and the arithmetic coding process for the coded pixel is performed. The configuration is based on the selected probability table.
  • the probability corresponding to the arithmetic code for coding the pixel to be coded is determined based on the pixel value of the coded pixel and the predicted pixel value of the uncoded pixel. Since a table is selected and the pixel value is encoded for the pixel to be encoded, the probability table in the arithmetic encoding process is adaptively switched for each pixel, so that efficient encoding can be performed. .
  • the present invention (Claim 7) is the image encoding device according to Claim 1, wherein the pixel value of the coded pixel is decoded to generate a locally decoded pixel value.
  • the encoding means is irreversible with respect to the pixel value of the encoded pixel based on the local encoded pixel value of the encoded pixel and the pseudo pixel value of the unencoded pixel. It is configured to perform encoding processing.
  • the lossy encoding process is performed on the pixel value of the pixel to be encoded based on the decrypted pixel value obtained by decrypting the pixel value of the encoded pixel. Therefore, the pixel value of the peripheral pixel referred to in the irreversible encoding process on the pixel to be encoded is the same as the pixel value of the peripheral pixel referenced in the decoding process on the pixel to be decoded. This makes it possible for the decoding side to correctly decode the image coded signal obtained by irreversible coding based on the pixel values of the peripheral pixels.
  • the present invention is the image encoding device according to Claim 7, wherein the encoding means comprises a locally decoded pixel value of the encoded pixel and a non-encoded pixel value of the unencoded pixel.
  • a prediction value generator for generating a predicted pixel value for the pixel to be encoded based on the pseudo pixel value; encoding a difference value between the pixel value of the pixel to be encoded and the predicted pixel value of the pixel to be encoded; And an encoder that outputs the encoded difference value as an image encoded signal in block units.
  • a predicted pixel value of a pixel to be coded is generated from the pixel value of the coded pixel and the pseudo pixel value of the uncoded pixel. Since the difference value between the pixel value and the predicted pixel value is encoded, the encoding amount required for encoding the pixel value of the pixel to be encoded can be reduced, and the encoding efficiency can be improved.
  • the present invention is the image encoding device according to Claim 7 or 8, wherein the encoding means comprises: a local decoded pixel value of the encoded pixel; Based on the pseudo pixel value of the uncoded pixel, the In this configuration, a code word for encoding a raw pixel value is selected.
  • a code for coding the pixel value of the coded pixel based on the pixel value of the coded pixel and the pseudo pixel value of the uncoded pixel. Since a word is selected, an efficient coding process can be realized by adaptively changing the code word for each pixel during the coding process of the image signal.
  • the encoding means comprises: encoding a pixel value of the encoded pixel; Based on the pseudo pixel value of the coded pixel, a probability table corresponding to a code code for arithmetically coding the pixel value of the coded pixel is selected, and the arithmetic coding process for the coded pixel is performed. The configuration is performed based on the selected probability table.
  • the probability corresponding to the arithmetic code for coding the pixel to be coded is determined based on the pixel value of the coded pixel and the pseudo pixel value of the uncoded pixel. Since a table is selected and the pixel value is encoded for the pixel to be encoded, the probability table in the arithmetic encoding process is adaptively switched for each pixel, so that efficient encoding can be performed. .
  • An image decoding apparatus sequentially converts pixel values constituting an image signal based on pixel values of a plurality of peripheral pixels located around a pixel to be encoded.
  • An image decoding device that performs an encoding process for each block of a predetermined number of pixels and decodes an encoded image signal obtained for each block.
  • a value substitution unit which receives an image coded signal composed of a plurality of pixel values corresponding to each of the blocks, and calculates each pixel value based on the pixel value of the decoded pixel and the pseudo pixel value of the undecoded pixel.
  • Decoding is performed in block units as described above.
  • the pixel value of the undecoded pixel among the plurality of peripheral pixels located around the pixel to be decoded is determined by the decoded pixel of the plurality of peripheral pixels.
  • the adaptive decryption processing in pixel units and the decryption processing in block units are combined without impairing the correlation of the pixel values between the undecoded pixel and the decoded pixel.
  • a decryption method can be realized. This has the effect of correctly decoding an image coded signal coded by a coding method combining adaptive coding processing in pixel units and coding processing in block units.
  • the present invention (Claim 12) is the image decoding device according to Claim 11, wherein the pixel value replacement means is a compound code having a shortest spatial distance from the undecoded pixel.
  • the pixel value of the decoded pixel is set as the pseudo pixel value of the undecoded pixel.
  • the pixel value of the decoded pixel closest to the undecoded pixel is used as the pseudo pixel value of the undecoded pixel.
  • the present invention (Claim 13) is the image decoding device according to Claim 11, wherein the pixel value replacing means is arranged so that a spatial distance from the undecoded pixel is the shortest.
  • the pixel value of the decrypted pixel located on the same horizontal scanning line as the undecoded pixel is set as the pseudo pixel value of the undecoded pixel.
  • the closest decoding pixel value to the above-mentioned undecoded pixel on the same horizontal scanning line as the undecoded pixel is used as the pseudo pixel value of the undecoded pixel. Since the pixel value of the decoded pixel is used, the pseudo pixel value of the undecoded pixel can be obtained by a simple method of holding the pixel value of the decoded pixel for a predetermined time.
  • the present invention is the image decoding device according to Claim 11, wherein the decoding means comprises a decoding unit that decodes the pixel value of the decoded pixel and the pseudo pixel of the undecoded pixel.
  • the decoding means comprises a decoding unit that decodes the pixel value of the decoded pixel and the pseudo pixel of the undecoded pixel.
  • a predicted value generator for generating a predicted pixel value for the pixel to be decoded based on the pixel value; a pixel value of the pixel to be decoded is decoded to generate a decoded pixel value;
  • a decoder for adding the predicted pixel value of the pixel to be decoded.
  • a predicted pixel value of a pixel to be decoded is generated from the pixel value of the decrypted pixel and the pseudo pixel value of the undecoded pixel, and the decoded pixel value is decoded. Since the predicted pixel value is added to the pixel value of the pixel to be decoded, the image coded signal obtained by encoding the difference between the pixel value of the pixel to be coded and the predicted pixel value is correctly decoded.
  • the present invention (claim 15) is the image decoding device according to claim 11 or claim 14, wherein the decoding means comprises: a pixel value of the decoded pixel; A code word for decoding the pixel value of the pixel to be decoded is selected based on the pseudo pixel value of the undecoded pixel.
  • the pixel value of the pixel to be decoded is decoded based on the pixel value of the decoded pixel and the pseudo pixel value of the undecoded pixel. Since the code word is selected, it is possible to correctly decode an image coded signal that has been subjected to efficient coding processing by adaptively changing the code word for each pixel.
  • the present invention is the image decoding apparatus according to Claim 11 or 14, wherein the decoding means comprises: Selecting a probability table corresponding to a code code for arithmetically decoding the pixel value of the decoded pixel based on the prime value and the pseudo pixel value of the undecoded pixel; The arithmetic decoding process is performed based on the selected probability table. According to the image decoding device having such a configuration, it corresponds to an arithmetic code for decoding a pixel to be decoded based on the pixel value of the decoded pixel and the pseudo pixel value of the undecoded pixel.
  • the image encoding method according to the present invention (claim 17) is a method for sequentially encoding pixel values constituting an image signal based on pixel values of a plurality of peripheral pixels located around a pixel to be encoded.
  • This is an image encoding method in which encoding processing for encoding is performed for each block composed of a predetermined number of pixels.
  • the pixel value is set as a reference pixel value.
  • a pseudo pixel value obtained based on a pixel value of an encoded pixel among the peripheral pixels based on a predetermined rule is set as a reference pixel value;
  • a pixel value of the to-be-encoded pixel is encoded based on reference pixel values set for a plurality of peripheral pixels with respect to the to-be-encoded pixel, and an image encoded signal for the image signal is generated. is there.
  • the image signal is blocked so as to correspond to each block on one image display area, and the pixel value of the pixel to be coded in the block is set to the pixel value of the peripheral pixel.
  • the peripheral pixel is a coded pixel
  • the pixel value is referred to.
  • the peripheral pixel is an uncoded pixel
  • the coded pixel is used instead of the pixel value. Since the pseudo pixel value obtained from the pixel value of the reference is referred to, in the above-described encoding process, the pixel to be encoded is located adjacent to the boundary of the block, and is not included in a plurality of peripheral pixels referred to.
  • a pseudo pixel value correlated with the pixel values of other peripheral pixels can be referred to as the pixel value of the uncoded pixel. Since this pseudo pixel value is obtained from the pixel values of the coded pixels around the coded pixel, the decoding side also refers to the undecoded pixel which is referred to when decoding the coded pixel. For, the pseudo pixel value obtained from the pixel value of the decrypted pixel can be referenced instead of the pixel value.
  • adaptive coding processing in pixel units and coding processing in block units are performed. Are combined without losing the correlation of the pixel values between the uncoded pixel and the pixel to be coded, while avoiding the difficulty of decoding the coded signal on the decoding side. be able to.
  • the pixel values constituting the image signal are sequentially determined based on the pixel values of a plurality of peripheral pixels located around the pixel to be encoded.
  • An image decoding method for decoding an image-encoded signal obtained by performing an encoding process for each block including a predetermined number of pixels for each block the image decoding method comprising: For a decoded pixel, its pixel value is set as a reference pixel value, and for an undecoded pixel of the plurality of peripheral pixels, a predetermined value is determined from a pixel value of a decoded pixel of the peripheral pixels.
  • a pseudo pixel value obtained based on the rule is set as a reference pixel value, and the pixel of the decoded pixel is set based on the reference pixel values set for a plurality of peripheral pixels with respect to the decoded pixel. Decode the value and And generates an image decoding signals for.
  • the image decoding method having such a configuration, for each block including a plurality of pixels, a process of sequentially decoding the pixel values of the pixels to be decoded in each block with reference to the pixel values of the peripheral pixels is performed.
  • the peripheral pixel is a decrypted pixel
  • the pixel value is referred to.
  • the peripheral pixel is an undecoded pixel
  • the pixel value is obtained from the pixel value of the decrypted pixel instead of the pixel value. Since the pseudo pixel value to be referred to is referred to is referred to, the decoding target pixel is located adjacent to the block boundary in the decoding process, and the undecoded pixel is located in a plurality of peripheral pixels referred to. Even if it is included, a pseudo pixel value that has a correlation with the pixel values of other peripheral pixels can be referred to as the pixel value of the undecoded pixel.
  • the adaptive decryption processing on a pixel basis and the decryption processing on a block basis impair the correlation of the pixel values between the undecrypted pixel and the decrypted pixel.
  • a combined decoding method can be realized.
  • the effect of being able to correctly decode an image-encoded signal encoded by an encoding method that combines adaptive encoding processing in pixel units and encoding processing in block units is obtained.
  • a data storage medium according to the present invention (claim 19) is a data storage medium storing a program for performing an encoding process or a decoding process on an image signal.
  • An image processing program for causing a computer to perform image signal processing by the image encoding method described in Item 17 or the image decoding method described in Item 18.
  • FIG. 1 is a block diagram showing a configuration of an image encoding device according to Embodiment 1 of the present invention.
  • FIG. 2 is a flowchart showing a process of generating a pixel value of an uncoded pixel in the encoding process by the image encoding device.
  • FIG. 3 is a block diagram showing a configuration of an image encoding device according to Embodiment 2 of the present invention.
  • FIG. 4 is a block diagram showing a configuration of an image encoding device according to Embodiment 3 of the present invention.
  • FIG. 5 is a block diagram showing a configuration of an image encoding device according to Embodiment 4 of the present invention.
  • FIG. 6 is a block diagram showing a configuration of an image encoding device according to a first modification of the fourth embodiment.
  • FIG. 7 is a block diagram showing a configuration of an image encoding device according to a second modification of the fourth embodiment.
  • FIG. 8 is a block diagram showing a configuration of an image decoding apparatus according to Embodiment 5 of the present invention.
  • FIG. 9 is a flowchart illustrating a process of generating a pixel value of an undecoded pixel in the decoding process by the image decoding device.
  • FIG. 10 is a block diagram showing a configuration of an image decoding apparatus according to Embodiment 6 of the present invention.
  • FIG. 11 is a block diagram showing a configuration of an image decoding apparatus according to Embodiment 7 of the present invention.
  • FIGS. 12 (a), 12 (b), and 12 (c) illustrate the encoding processing by the image encoding apparatus or the decoding processing by the image decoding apparatus according to the above embodiment.
  • FIG. 2 is a diagram showing a data storage medium storing a program to be realized by a computer system.
  • FIG. 13 (a) and FIG. 13 (b) are schematic diagrams showing a manner in which one frame screen is divided into a plurality of blocks in the encoding process in block units.
  • FIG. 14 is a schematic diagram for explaining adaptive encoding processing in pixel units.
  • FIG. 15 is a schematic diagram for explaining a problem in a combination encoding method that combines the above-described encoding processing in units of blocks and encoding processing in units of pixels.
  • FIG. 16 (a) and FIG. 16 (b) are schematic diagrams showing a manner of dividing one frame screen into a plurality of blocks in the block-based decoding processing.
  • FIG. 17 is a schematic diagram for explaining adaptive decoding processing in pixel units.
  • FIG. 18 is a schematic diagram for explaining a problem in a combination decoding method in which the above-described decoding processing in block units and decoding processing in pixel units are combined.
  • FIG. 1 is a block diagram showing a configuration of an image encoding device 101 according to Embodiment 1 of the present invention.
  • the image encoding device 101 includes a block generator 2 that divides an input image signal Is to correspond to each of a plurality of blocks constituting one image display area (one frame); Receiving the output Bs of the encoder 2 and calculating the pixel value of the coded pixel to be coded in the coded block to be coded, by the predicted pixel value of the coded pixel (hereinafter simply referred to as the predicted value It has an encoder 16a that performs lossless encoding with reference to Sp, and a predicted value generator 110 that generates the predicted pixel value Sp.
  • the predicted value generation unit 110 includes a large-capacity main memory 4 capable of storing the pixel values of each pixel constituting the input image signal by, for example, a number corresponding to one frame, and outputting from the main memory 4
  • the first and second auxiliary memories 6a and 6b having different holding times for temporarily holding the pixel values M to be stored are provided.
  • the main memory 4 stores the reference pixel P 0 of the coded pixel PX to be processed by the encoder 16 a during the time required for the encoding process of one pixel value.
  • PP 9 (see FIG. 14) are sequentially output from the stored pixel values.
  • the first auxiliary memory 6a is configured to delay the pixel value M sequentially output from the main memory 4 by one pixel, and the second auxiliary memory 6b is sequentially output from the main memory 4.
  • the pixel value M is delayed by two pixels.
  • the predicted value generation unit 110 receives the image input signal Is of each frame, counts the number of pixel values, and outputs an output C out of the counter 8 and each frame supplied from the outside. Whether the pixel value output from the main memory 4 is the pixel value of a coded pixel that has already been coded by the coder 16a based on the information BNs of the number of blocks in the horizontal and vertical directions , Still not encoded And a coding / uncoding judging unit 10 for judging whether the pixel value is a pixel value of an uncoded pixel that has not been set.
  • the determiner 10 also performs a process of measuring the distance to the closest coded pixel located on the same horizontal scanning line as the number of pixels.
  • the counter 8 is reset when the pixel values of all the pixels constituting one frame are input.
  • the predicted value generation unit 110 outputs one of the main memory 4, the first and second auxiliary memories 6a and 6b based on the output of the coded / uncoded determinator 10
  • a selection switch 12 for selecting and outputting M, Ma, and Mb, and an output S out of the selection switch 12 are used as reference pixels necessary for generating a predicted pixel value corresponding to the pixel to be encoded PX.
  • a prediction value generator 14 that obtains pixel values P0 to P9 and generates a prediction pixel value Sp corresponding to the pixel to be encoded PX.
  • the encoder 16a calculates the difference between the pixel value of the pixel to be encoded PX output from the block encoder 2 and its predicted pixel value.
  • the encoded difference value is output as an encoded pixel value corresponding to the above-mentioned encoded pixel PX.
  • a codeword for encoding the pixel value of the pixel to be encoded PX is selected based on the predicted pixel value obtained from the pixel values of the reference pixels P0 to P9. ing.
  • FIG. 2 is a flowchart showing the encoding process of the image encoding device 101. First, the flow of the encoding process will be briefly described according to the flowchart.
  • step S1 When an image signal is input to the present image encoding device 101, pixel values constituting the image signal are sequentially stored in the main memory 4, and from the main memory 4, encoding of the pixel PX to be encoded is performed.
  • step S2 the first reference pixel among the plurality of reference pixels acquired in the main memory 4 is determined as a determination target pixel (step S2), and the determination target pixel is encoded.
  • the determination as to whether or not they are prime is made by the coding / uncoding determining unit 10 (step S3).
  • the coded / uncoded determinator 10 sequentially increments the counter output C out and the pixel values of the nine reference pixels P 0 to P 9 from the main memory 4 based on the reference clock. Since the operation is synchronized, the position of the reference pixel output from the main memory 4 with respect to the pixel to be encoded PX can be detected from the counter output Cout.
  • the reference pixel to be determined is a coded pixel
  • the pixel value is obtained as the reference pixel value by the prediction value generator 14 (Step S5)
  • a reference pixel value (pseudo pixel value) is generated from the encoded pixels around the reference pixel, and the prediction value generator Acquired at 14 (step S 4).
  • step S6 it is determined whether or not all the reference pixel values necessary for encoding the pixel to be encoded have been obtained by the predicted value generator 14 (step S6), and the necessary reference pixel value is determined. If all the reference pixels have not been acquired, the next reference pixel among the reference pixels acquired in the main memory 4 is set as the determination target (step S8), and the processing in steps S3 to S6 is repeatedly performed. Will be On the other hand, if all of the necessary reference pixel values have been acquired by the predicted value generator 14, the predicted value generator 14 calculates the coded pixel PX based on the reference pixel values. Is generated (step S7).
  • step S9 the predicted pixel value of the coded pixel PX is obtained from the predicted value generator 14 at the same time.
  • an encoding process for the pixel value of the encoded pixel PX is performed using the predicted pixel value (step S10).
  • a plurality of pixels constituting the image input signal Is constitute one frame by the blocker 2.
  • pixel values corresponding to the pixels in each block are sent to an encoder 16a, and the encoder 16a Encoding processing for encoding the pixel value of the pixel PX with reference to the reference pixel value for each pixel is performed in units of blocks.
  • the main memory 4 sequentially stores the pixel values constituting the image input signal Is having the scanning line structure, and the pixel values of the reference pixels P0 to P9 corresponding to the pixel PX to be encoded are
  • the data is output at a constant read cycle (step S1).
  • the output M of the main memory 4 is temporarily stored in the first and second auxiliary memories 6a and 6b.
  • the output of the main memory 4 is held only for one read cycle, and in the second auxiliary memory 6b, it is held for two read cycles.
  • the counter 8 counts the number of pixel values input based on the image input signal Is with reference to the first pixel value of one frame, and this count value Cout is encoded / unprocessed. It is output to the encoding determiner 10.
  • the coded Z uncoded determiner 10 receives the vertical and horizontal block count information BN s in one frame from the outside, and the block count information BN s and the reference designated by the output of the counter 8 described above.
  • the pixel is a determination target pixel for determining whether the pixel is the coded pixel or the uncoded pixel. For example, when the pixel to be coded PX is determined as shown in FIG. 14, the corresponding reference pixels P0 to P9 are determined, and the pixel P from which the pixel value is first output from the main memory 4 is determined. 0 is the pixel to be judged first (step S 2).
  • the pixel value of the reference pixel P0 is held in the first and second auxiliary memories 6a and 6b for one reading period and two reading periods, respectively.
  • the coded / uncoded determinator 10 calculates the position of the coded pixel in the coded block based on the counter output C 0 ut and the block number information BN s, and It is determined whether the pixel is a coded pixel or an uncoded pixel (step S3), and the selection switch 12 is controlled according to the result of the determination. Since the reference pixel P 0 is a coded pixel as shown in FIG.
  • the selection switch 12 is / Output M of the main memory 4 is selected by the uncoded judgment unit 10.
  • the pixel value of the reference pixel P0 is stored in the prediction value generator 14 as a reference pixel value.
  • the encoding / unencoding determination unit 10 determines whether or not the pixel values of all reference pixels corresponding to the pixel to be encoded Px have been acquired by the predicted value generator 14. (Step S6). In this case, since all the reference pixel values of the reference pixels P0 to P9 have not been acquired by the prediction value generator 14, the coded Z Then, the pixel value of the reference pixel P1 output next to the pixel value of the reference pixel P0 is set as the pixel value of the determination target pixel (step S8). Since the reference pixel P1 is a coded pixel in the same manner as the reference pixel P0, the processing in steps S3, S5, S6, and S8 is performed on the pixel value of the reference pixel P1. Will be
  • the pixel value of the reference pixel P2 output next to the pixel value of the reference pixel P1 from the main memory 4 is set as the pixel value of the determination target pixel.
  • the selection switch 12 is set to the coded Z uncoded.
  • the determination unit 10 is controlled to select the output Ma of the first auxiliary memory 6a, whereby the pixel value of the reference pixel P1 becomes a pseudo-pixel value of the reference pixel P2. 4 (step S 4). Thereafter, the processing of the above-described steps S6 and S8 is performed.
  • steps S3, S5, S6, and S8 are performed for the reference pixels P3 to P5 in the same manner as the above-described reference pixel P0.
  • the processing of steps S3, S4, S6, and S8 is performed.
  • the pixel value of the reference pixel P7 output next to the pixel value of the reference pixel P6 from the main memory 4 is set as the pixel value of the determination target pixel.
  • the reference pixel P7 is an uncoded pixel located one pixel away from the coded pixel PX.
  • Uncoded judgment unit Control is performed to select the output Mb of the second auxiliary memory 6b, whereby the pixel value of the reference pixel p5 is stored in the prediction value generator 14 as a pseudo pixel value of the reference pixel P7.
  • the above-described steps S 3, S 5, and S 6 are performed on the reference pixels P 8 and P 9 in the same manner as the above-described reference pixel P 0.
  • the encoding / uncoding unit 10 determines that the reference pixel values for all of the reference pixels PO to P9 have been acquired by the prediction value generator 14, and the prediction value The generator 14 calculates a predicted pixel value Sp corresponding to the pixel to be encoded PX based on the acquired reference pixel value (step S7).
  • the encoder 16a obtains the predicted pixel value Sp along with the pixel value of the pixel to be encoded PX (step S9), and the pixel value of the pixel to be encoded PX and its predicted pixel value S
  • An encoding process is performed on the difference value from p, and the encoded difference value is output as an encoded signal of the above-mentioned encoded pixel PX (step S10).
  • a codeword selected based on the predicted pixel value Sp is used.
  • the pixel values of each block in one frame are sequentially encoded. Note that the encoding process is performed with the predicted pixel value being 0, since there is no encoded reference pixel around the head pixel of one frame.
  • the processing and the encoding processing in units of blocks can be performed without impairing the correlation between the pixel values of the uncoded pixels and the pixels to be coded, and without making it difficult to decode the coded signal. You can combine them. This makes it possible to converge the effects of transmission errors on a block-by-block basis and to improve coding efficiency compared to simple block-by-block coding.In addition, the decoding side predicts the pixel to be coded.
  • the decoding process on the encoded signal without causing the deterioration in efficiency.
  • the decoding process on the pixel to be decoded is performed on the decoding side as well. This is done by referring to the value, and it is necessary to match the pixel value referenced during the encoding process with the pixel value referenced during the decoding process.
  • a pseudo pixel value for the uncoded pixel is generated from the referenced coded pixel by a predetermined rule.
  • the pixel values of the uncoded pixels P0, P1, P3, P4, P5, P8, and P9 are calculated. It is uniquely determined from the pixel value.
  • the pixel value of the coded pixel located on the same horizontal scanning line as the uncoded pixel and having a short distance is set as the pseudo pixel value for the uncoded pixel.
  • the pixel value of the uncoded pixel P2 is replaced by the pixel value of the coded pixel P1
  • the pixel values of the uncoded pixels P6 and P7 are replaced by the pixel value of the coded pixel P5. Replaced by the pixel value.
  • the pixel value correlation between the coded pixel and the uncoded pixel at the block boundary becomes large, and highly efficient coding is performed using the same prediction method as inside the block where the pixel correlation is large. Processing can be realized.
  • auxiliary memories 6a and 6b having different times for holding the output M of the main memory 4 are provided.
  • the auxiliary memory for temporarily holding the output of the main memory 4 is provided.
  • the time for holding the pixel value output from the main memory 4 is switched by the coded / uncoded determinator 10 according to the distance of the uncoded pixel from the coded pixel. You may.
  • the prediction of the coded pixel PX is performed.
  • the encoding process may be performed based not only on the pixel value but also on the prediction probability indicating the likelihood of the predicted pixel value.
  • the predicted pixel value is either “0” or “1”. It may be fixed and refer to only the predicted probability.
  • Embodiment 2 of the present invention an image coding apparatus that performs a coding process of a binary shape signal with reference to only the prediction probability instead of the prediction pixel value, and as Embodiment 3, With reference to the prediction probability together with the prediction pixel value, each of the image encoding devices that performs the encoding process of the multi-level image signal will be described.
  • FIG. 3 is a block diagram showing a configuration of an image encoding device 102 according to Embodiment 2 of the present invention.
  • the same reference numerals as those in FIG. 1 denote the same parts as those in the image coding apparatus 101 of the first embodiment.
  • This image coding device 102 includes a prediction probability generator 22 instead of the prediction value generator 14 in the image coding device 101 of the first embodiment, and has pixel values of “0” and “ 1 ".
  • the prediction probability generator 22 calculates the probability that the pixel value of the pixel to be encoded PX and its predicted pixel value match from the pixel values of the reference pixels P0 to P9 with respect to the pixel to be encoded PX as the prediction probability. Then, the prediction probability signal S k for the pixel to be encoded PX is output to the encoder 16b.
  • the encoder 16b subtracts the pixel value of the pixel to be encoded PX from the pixel value of the pixel to be encoded PX because the image encoding apparatus 102 is intended for encoding of the binary shape signal. Is set to one of “0” and “1”. When the prediction probability is large, the encoder 16b has a high probability that the pixel value of the pixel to be coded PX and the predicted pixel value match. And the difference between the prediction pixel value and the prediction pixel value are encoded by an encoding method that increases the encoding efficiency when the difference value becomes 0.
  • the coding is performed by a coding method that increases the coding efficiency.
  • the pixel values are generated from the pixel values of the coded pixels in the same manner as in the image coding apparatus 101 of the first embodiment, and the pixel values corresponding to all the reference pixels P0 to P9 are used to generate the prediction probability. It is stored in the container 22.
  • the prediction probability generator 22 calculates a prediction probability for the coded pixel PX based on the pixel values of the reference pixels P0 to P9.
  • the encoder 16b presets the pixel value of the pixel PX to be encoded.
  • the coding process according to the above-mentioned prediction probability information S k is performed on the difference value from the predicted pixel value.
  • the image coding apparatus 102 that performs the coding process of the binary shape signal, the coded pixels adjacent to the block boundary and having difficulty in generating the predicted value are coded. Encoding can be performed while suppressing deterioration in efficiency.
  • the encoding process based on the prediction probability is described in, for example, the international standard JBIG (Joint Bi-level Image Coding Experts Group). Since the encoding process is performed in pixel units (non-block units), the reference pixels are always encoded pixels. In the encoding process in block units, which is a problem to be solved by the present invention, the reference pixels are not encoded. There is no disclosure of how to set the pixel value for this uncoded pixel, as well as the problem in the case of a pixel.
  • JBIG Joint Bi-level Image Coding Experts Group
  • FIG. 4 is a block diagram showing a configuration of an image encoding device 103 according to Embodiment 3 of the present invention.
  • the same reference numerals as those in FIG. 1 denote the same parts as those in the image coding apparatus 101 of the first embodiment.
  • the image encoding device 103 performs encoding processing on a multi-valued image signal as described above, and the prediction probability generation unit 130 performs the image encoding according to the first embodiment.
  • This is a configuration having a prediction probability generator 22 in addition to the prediction value generator 14 in the device 101.
  • the prediction probability generator 22 receives the output S out of the selection switch 12 similarly to the prediction value generator 14 and sequentially stores the pixel values of the reference pixels P 0 to P 9 with respect to the coded pixel PX.
  • the encoder 16 c calculates the pixel value of the pixel to be encoded PX. Is encoded.
  • the encoding process is performed on the pixel value of the pixel to be encoded PX by an encoding method.
  • the encoder 16c calculates the difference value between the pixel value and the predicted pixel value. If the value is slightly larger, the encoding process is performed on the pixel value of the pixel to be encoded PX by an encoding method with high encoding efficiency.
  • the predicted value generator 14 obtains the predicted pixel value of the coded pixel PX from the reference pixels P0 to P9 in the same manner as in the first embodiment.
  • the prediction probability generator 22 calculates a prediction probability S k for the pixel to be encoded PX based on the pixel values of the reference pixels P 0 to P 9.
  • the encoder 16c When the predicted pixel value Sp and the predicted probability Sk are output to the encoder 16c, the encoder 16c outputs the pixel value of the pixel to be encoded PX and the predicted value from the predicted value generator 14. An encoding process according to the prediction probability S k is performed on the difference value from the predicted pixel value.
  • a pixel to be encoded adjacent to a block boundary and for which it is difficult to generate a predicted value is degraded in encoding efficiency.
  • Can be encoded while suppressing As a result, the effects of transmission errors can be converged on a block-by-block basis, and the coding efficiency can be improved as compared with simple block-based coding processing. Decoding of the encoded signal can be performed correctly without any deterioration in efficiency.
  • FIG. 5 is a block diagram showing a configuration of an image encoding device 104 according to Embodiment 4 of the present invention.
  • the same reference numerals as those in FIG. 1 denote the same parts as those in the image coding apparatus 101 of the first embodiment.
  • This image coding apparatus 104 is different from the image coding apparatus 101 of the first embodiment in that a lossy coding process is performed on the blocked image signal Bs.
  • the image encoding device 104 replaces the encoder 16a that performs the lossless encoding process of the first embodiment with a DCT (discrete) for the output Bs of the block encoder 2.
  • the encoder 16 d performs irreversible encoding including cosine transform based on the predicted pixel value from the predicted value generator 14.
  • the predicted value generation unit 140 of the image coding device 104 decodes the output C s of the encoder 16 d based on the predicted pixel value Sp from the predicted value generator 14.
  • a local decoder 24; an output LDs of the local decoder 24 is stored in the main memory 4 as a pixel value of a decoded pixel; and the output LDs is stored in the counter 8 To be entered.
  • Other configurations are the same as those of the image coding apparatus 101 of the first embodiment.
  • the local decoder 24 decodes the output Cs of the encoder 16 d with reference to the predicted pixel value Sp, and stores the decoded pixel value in the main memory 4. To be stored.
  • the decoded pixel value is used to generate the predicted pixel value, and thus the image encoded by the image encoding device is encoded. It becomes possible to correctly decode the coded image signal by the image decoding device.
  • the predicted pixel value is either “0” or “1”. It may be fixed to one side and refer to only the prediction probability.
  • an image coding apparatus that performs irreversible coding of a binary shape signal by referring to only the prediction probability instead of the above-described predicted pixel value is described.
  • an image encoding device that performs irreversible encoding of a multi-valued image signal with reference to the prediction probability together with the predicted pixel value will be described.
  • FIG. 6 shows an image coding apparatus 104 a according to a first modification of the fourth embodiment of the present invention.
  • FIG. 3 is a block diagram showing the configuration of FIG. Note that, in the figure, the same reference numerals as those in FIG. 5 denote the same parts as those in the image coding apparatus 104 according to the fourth embodiment.
  • This image coding device 104 a includes a prediction probability generator 22 instead of the prediction value generator 14 in the image coding device 104 of the fourth embodiment, and has a pixel value of It performs encoding processing of a binary shape signal that is either “0” or “1”.
  • the prediction probability generator 22 calculates the probability that the pixel value of the pixel to be encoded PX and its predicted pixel value match from the pixel values of the reference pixels P0 to P9 with respect to the pixel to be encoded PX as the prediction probability. Then, the prediction probability S k for the pixel to be encoded PX is output to the encoder 16 e and the local decoder 24.
  • the encoder 16e since the present image encoding apparatus processes a binary shape signal, the predicted pixel values for the pixel to be encoded P x are “0” and “1”. Is set to one of Further, like the encoder 16b according to the second embodiment, when the prediction probability is large, the encoder 16e calculates the pixel value of the encoded pixel PX and its predicted pixel value. If the prediction value is small, the pixel value of the pixel to be coded PX and the prediction value thereof are encoded if the prediction value is small. The difference value from the pixel value is encoded by an encoding method that increases the encoding efficiency when the difference value is not 0.
  • the local decryption unit 24 switches the decryption method based on the prediction probability and performs the decryption process in the same manner as the encoder 16e.
  • the description of the operations common to those of the image coding apparatus according to the fourth embodiment is omitted.
  • the prediction probability generator 22 calculates the pixel values of the reference pixels P0 to P9. , The prediction probability S k for the pixel to be coded P x is obtained.
  • the encoder 16 e outputs the pixel of the pixel PX to be encoded.
  • An irreversible encoding process according to the prediction probability S k is performed on a difference value between the value and a preset prediction pixel value.
  • the local decoder 24 performs a decoding process on the output C s of the encoder 16 e according to the prediction probability S k, and obtains a pixel value of the pixel PX to be encoded.
  • the reproduced pixel value of the encoded pixel PX is stored in the main memory 4.
  • FIG. 7 is a block diagram showing a configuration of an image coding device 104 b according to a second modification of the fourth embodiment of the present invention. 4, the same reference numerals as those in FIG. 4 denote the same components as those in the image coding apparatus 104 according to the fourth embodiment.
  • the image encoding device 104b performs irreversible encoding processing on a multi-valued image signal as described above, and the prediction probability generation unit 140b is used in the fourth embodiment.
  • a prediction probability generator 22 is provided in addition to the prediction value generator 14 in the image encoding device 104 of FIG.
  • the prediction probability generator 22 receives the output S 0 ut of the selection switch 12 similarly to the prediction value generator 14 and sequentially stores the pixel values of the reference pixels P 0 to P 9 for the pixel to be encoded PX. Based on these pixel values, the prediction value generator 14 outputs a prediction probability S k indicating the likelihood of a prediction pixel value Sp predicted based on the reference pixels P 0 to P 9. It has become.
  • the encoder 16 f calculates the pixel to be coded based on the prediction pixel value Sp from the prediction value generator 14 and the prediction probability S k from the prediction probability generator 22.
  • the configuration is such that a lossy encoding process is performed on the pixel value of Px, and the specific configuration is exactly the same as the encoder 16c of the third embodiment.
  • the local decoder 2 switches the encoding method based on the prediction probability and performs a decoding process in the same manner as the encoder 16e. In the encoding processing of the multi-valued image signal by the image encoding device 104 b having such a configuration, even the reference pixels P 0 to P 9 corresponding to the encoded pixel PX correspond to the uncoded pixels.
  • the predicted value generator 14 obtains a predicted pixel value of the coded pixel PX from the reference pixels P0 to P9 in the same manner as in the fourth embodiment.
  • the prediction probability generator 22 calculates a prediction probability S k for the pixel to be encoded PX based on the pixel values of the reference pixels P 0 to P 9.
  • the encoder 16 f calculates the pixel value of the pixel to be encoded PX in advance.
  • An encoding process according to the above-mentioned prediction probability S k is performed on the difference value from the set prediction pixel value.
  • the local decoder 24 performs a decoding process on the output C s of the encoder 16 f according to the prediction probability S k to calculate the pixel value of the pixel to be encoded PX. Reproduce.
  • the reproduced pixel value of the encoded pixel PX is stored in the main memory 4.
  • the coding efficiency is not degraded for an image signal for which prediction value generation is difficult, and a large improvement in coding efficiency for an image signal for which prediction value generation is easy can be achieved.
  • the image encoding device 104 b that lossily encodes a multilevel image signal, it is difficult to generate a predicted value adjacent to a block boundary. Pixels can be coded while suppressing deterioration in coding efficiency. For this reason, the effects of transmission errors can be converged on a block-by-block basis, and the coding efficiency can be improved compared to simple block-based coding processing. In addition, on the decoding side, the decoding efficiency of the encoded signal can be correctly performed without deterioration in the prediction efficiency of the pixel to be encoded.
  • FIG. 8 is a block diagram showing a configuration of an image decoding apparatus 105 according to Embodiment 5 of the present invention.
  • An image decoding apparatus 105 according to the fifth embodiment is configured to reversibly decode an image coded signal that has been subjected to the lossless encoding process by the image coding apparatus 101 according to the first embodiment. is there.
  • the image decoding apparatus 105 decodes the input image coded signal Cs for each of a plurality of blocks constituting one image display area.
  • the decoding target to be decoded is A decoder 26a for decoding the pixel value of the coded pixel P ⁇ ', that is, the coded signal obtained by coding the pixel value of the coded pixel PX, based on the predicted pixel value of the coded pixel;
  • a deblocking unit 30 that integrates the image decoding signal Ds corresponding to each block output from the decoder 26a to generate an image reproduction signal Rs having a predetermined scanning line structure;
  • a predicted value generation unit 150 that generates the predicted pixel value based on the pixel values of reference pixels P 0 ′ to ⁇ 9 ′ located around the decoded pixel P x ′.
  • the predicted value generation unit 150 has substantially the same configuration as the predicted value generation unit 110 in the image encoding device 101 of the first embodiment.
  • the predicted value generation unit 150 includes, for example, a large-capacity main memory 4 capable of storing a number corresponding to one frame, and a holding time for temporarily holding the pixel value ⁇ output from the main memory 4. And first and second auxiliary memories 6a and 6b.
  • the main memory 4 stores the reference pixel of the pixel to be decoded ⁇ ⁇ ⁇ ⁇ ′ being processed by the decoder 26 a during the time required for the encoding process of one pixel value.
  • the pixel values corresponding to ⁇ ⁇ 'to ⁇ 9' are sequentially output from the stored pixel values.
  • the first auxiliary memory 6a is configured to delay the pixel value M sequentially output from the main memory 4 by one pixel
  • the second auxiliary memory 6b is sequentially output from the main memory 4.
  • Pixel value M is delayed by two pixels.
  • the predicted value generation unit 150 receives the image decoded signal Ds of each frame, counts the number of pixel values, and outputs an output C out of the counter 8 and each of the counters supplied from the outside. Based on the information BN s of the number of blocks in the frame in the vertical and horizontal directions, the pixel value output from the main memory 4 is the pixel value of the decoded pixel that has already been subjected to the decoding processing by the encoder 26 a.
  • a decoding / undecoding determining unit 20 that determines whether the pixel value is the pixel value of an undecoded pixel that has not been subjected to the decoding process.
  • the determiner 20 also performs a process of measuring the distance to the nearest decoded pixel located on the same horizontal scanning line as the number of pixels.
  • the counter 8 is reset when the pixel values of all the pixels constituting one frame are input.
  • the predicted value generation unit 150 is configured to output any one of the main memory 4, the first and second auxiliary memories 6 a and 6 b based on the output S cont of the decoding / non-decoding determining unit 20.
  • a selection switch 12 for selecting and outputting the outputs M, M a, and M b of the pixel and an output S out of the selection switch 12 for generating a predicted pixel value corresponding to the pixel to be decoded P x ′
  • a prediction value generator 14 that obtains the pixel values of the reference pixels P 0 ′ to ⁇ 9 ′ and generates a prediction pixel value S ⁇ corresponding to the decrypted pixel ⁇ ⁇ ′.
  • the decoder 26a decodes the coding difference value for the coded pixel PX input as an image coding signal from the outside to decode the decoding difference value.
  • a value is generated, and the predicted pixel value Sp from the predicted value generation unit 150 is added to the decryption difference value to generate a decoded pixel value for the pixel to be decoded. Output.
  • FIG. 9 is a flowchart showing the decryption processing of the image decoding apparatus 105. First, the flow of the decryption processing will be briefly described according to this flowchart.
  • the coded image signal Cs is input to the present image coding device 105, and the decoding process of the coded image signal Cs is performed by the decoder 26a by the prediction value generation unit 150. This is performed based on the signal Sp.
  • a plurality of decrypted pixel values corresponding to one frame are sequentially stored in the main memory 4 as an output of the decryptor 26a, and from the main memory 4, the decryption of the decrypted pixel ⁇ ⁇ is performed.
  • the pixel values of a plurality of reference pixels (peripheral pixels) ⁇ ⁇ ′ to ⁇ 9 ′ which are located around the pixel to be decoded ⁇ ′ which are referred to at the time of encoding are output (step S11). .
  • the first reference pixel of the plurality of reference pixels acquired in the main memory 4 is determined as a determination target pixel (step S12), and whether or not this determination target pixel is a decoded pixel is determined.
  • the determination is performed by the decryption / non-decryption determination unit 30 (step S13).
  • the increment of the counter output Cout and the pixel values of the nine reference pixels P0 'to P9' from the main memory 4 are sequentially determined based on the reference clock. Since the output operation is synchronized, the position of the reference pixel output from the main memory 4 with respect to the decrypted pixel ⁇ ⁇ ′ can be detected from the counter output C out.
  • the reference pixel to be determined is a decoded pixel
  • the pixel value is obtained by the above-described predicted value generator 14 as a reference pixel value (step S15), and the above determination is made. If the reference pixel that is the target pixel is not a decoded pixel, a reference pixel value (pseudo-pixel value) is generated from the decoded pixels around the reference pixel, and the predicted value generator 14 (Step S14).
  • step S16 it is determined whether or not all the reference pixel values necessary for decoding the pixel to be decoded have been obtained by the prediction value generator 14 (step S16). If all the reference pixels have not been obtained, the next reference pixel among the reference pixels obtained in the main memory 4 is determined as the above-mentioned determination target (step S18), and the processing of steps S13 to S16 is repeated. Done. On the other hand, when all of the necessary reference pixel values have been obtained by the predicted value generator 14, the predicted value generator 14 sets the decoded pixel P x ′ based on the reference pixel values. A corresponding predicted pixel value is generated (step S17).
  • the pixel value of the pixel to be decoded P ⁇ ′ is obtained by the decoder 26a.
  • the predicted pixel value of the pixel to be decoded P x ′ is obtained from the predicted value generator 14 (step S19), and the pixel value of the pixel to be decoded P ⁇ ′ is decoded by the decoder 26a. Is performed using the above-mentioned predicted pixel value (step S19).
  • the decoded image signal Cs is sent to a decoder 26a, where the pixel to be decoded is The decoding process of decoding the pixel value of ⁇ ⁇ ′ for each pixel with reference to the reference pixel value is performed on a block basis.
  • the main memory 4 sequentially stores the pixel values constituting one frame, which is the output of the decoder 26a, and the reference pixel P 0 ′ corresponding to the pixel P x ′ to be decoded.
  • the output ⁇ of the main memory 4 is temporarily stored in the first and second auxiliary memories 6 a,
  • the output of the main memory 4 is held only for one read cycle, and in the second auxiliary memory 6b, it is held for two read cycles.
  • the number of input pixel values is counted based on the output D s of the decoder 26a based on the first pixel value of the one frame, and the count value C out Is output to the decryption / non-decoding determination unit 20.
  • the vertical and horizontal block number information BNs in one frame is input from the outside, and the reference pixel specified by the block number information BNs and the output of the counter 8 is used.
  • the pixel is a determination target pixel for determining whether the pixel is a decoded pixel or an undecoded pixel. For example, when the pixel to be decrypted P ⁇ ′ is determined as shown in FIG. 17, the corresponding reference pixels P 0 ′ to ⁇ 9 ′ are determined.
  • the pixel P0 'to be output is the pixel to be determined first (step S12).
  • the decryption Z non-decryption determiner 20 calculates the position of the pixel to be decrypted in the decrypted block based on the counter output C out and the block number information BNs. It is determined whether the pixel is a decrypted pixel or an undecoded pixel, and the selection switch 12 is controlled according to the determination result.
  • this pixel P 0 ′ is a decrypted pixel as shown in FIG. 17, as a result of the determination (step S 13) by the decrypted Z
  • the decoding / non-decoding determiner 20 is controlled to select the output M of the main memory 4, whereby the pixel value of the pixel P 0 ′ is stored in the prediction value generator 14 as a reference pixel value.
  • the decoding / non-decoding determining unit 20 determines whether or not the pixel values of all the reference pixels corresponding to the pixel to be decoded P x ′ have been acquired by the prediction value generator 14 (step S 16). In this case, since all the reference pixel values of the reference pixels P 0 ′ to ⁇ 9 ′ have not been acquired by the prediction value generator 14, the decoding Then, the pixel value of the reference pixel ⁇ 1 ′ output next to the pixel value of the reference pixel ⁇ 0 ′ is set as the pixel value of the pixel to be determined (step S18). Since the reference pixel P 1 ′ is a decoded pixel as in the case of the reference pixel P 0 ′, the pixel values of the reference pixel P 1 ′ are also determined in steps S 13, S 15, S 16, S Step 18 is performed.
  • the pixel value of the reference pixel ⁇ 2 ′ output next to the pixel value of the reference pixel P 1 ′ from the main memory 4 is used as the determination target pixel.
  • the pixel value is set (step S18).
  • the reference pixel ⁇ 2 ′ is an undecoded pixel located next to the decoded pixel.
  • the output Ma of the first auxiliary memory 6 a is controlled to be selected by the undecoding determiner 20, whereby the pixel value of the reference pixel P 1 ′ is changed to the pseudo pixel of the reference pixel P 2 ′. It is stored as a value in the predicted value generator 14 (step S14). After that, the processing of the above steps S16 and S18 is performed.
  • Steps SI3, SI5, SI6, and SI8 are performed in the same manner as described above, and for the reference pixel p6 ', steps S13', S1 ', S16 ', S18' are performed.
  • the pixel value of the reference pixel P 7 ′ output next to the pixel value of the reference pixel P 6 ′ from the main memory 4 is the pixel of the pixel to be determined.
  • the value is set (step S18).
  • the reference pixel is an undecoded pixel located one pixel away from the decoded pixel.
  • the undecoded judgment unit 20 controls to select the output Mb of the second auxiliary memory 6b, whereby the pixel value of the reference pixel P5 'is predicted as a pseudo pixel value of the reference pixel P7'. It is stored in the value generator 14 (step S14). Thereafter, the processing of the above-described steps S16 and S18 is performed.
  • steps S13, S15, and S16 are performed on the reference pixels P8 'and P9' in the same manner as the reference pixel P.
  • the decryption / non-decryption determination unit 20 determines that the reference pixel values for all of the reference pixels P 0 ′ to ⁇ 9 ′ have been acquired by the prediction value generator 14,
  • the predicted value generator 14 calculates a predicted pixel value corresponding to the pixel to be decoded ⁇ 'based on the acquired reference pixel value (step S17).
  • the above-mentioned predicted pixel value is obtained by the decoder 26a together with the pixel value of the pixel to be decoded ⁇ ⁇ ′ (step S19), and the pixel value of the pixel to be decoded P x ′ and its prediction are obtained.
  • the value added to the pixel value is output as the image decryption signal Ds of the decrypted pixel PX (step S20).
  • the pixel values of each block in one frame are decoded. Note that, for the first pixel of one frame, since there is no reference pixel subjected to the decoding processing, the decoding processing is performed with the predicted pixel value set to 0.
  • the image decryption signal Ds is integrated by the reverse blocker 30 so as to correspond to one frame screen, and an image reproduction signal Rs having a scanning line structure is output.
  • the pixel value of an undecoded pixel when the pixel value of an undecoded pixel is referred to, the pixel value of an encoded pixel in the vicinity is referred to.
  • Adaptive decoding processing and decoding processing in units of blocks can be performed without impairing the correlation of pixel values between undecoded pixels and decoded pixels. This makes it possible to converge the effects of transmission errors on a block-by-block basis, and correctly decodes the image coded signal Cs processed by a coding method that has improved coding efficiency compared to simple block-based coding processing.
  • the two auxiliary memories 6a and 6b having different times for holding the output M of the main memory 4 are provided.
  • the auxiliary memory for temporarily holding the output of the main memory 4 is provided.
  • the time to hold the pixel value output from the main memory 4 is determined by the decrypted Z undecoded determiner 20 according to the distance of the undecoded pixel from the pixel to be decoded. Switching may be performed.
  • Embodiment 5 when encoding the image signal, only the predicted pixel value of the pixel to be decoded P ⁇ ′ is referred to, but the prediction pixel of the pixel to be decoded ⁇ ⁇ ′ is referred to.
  • the decoding process may be performed based not only on the value but also on the prediction probability indicating the likelihood of the predicted pixel value.
  • the predicted pixel value is either “0” or “1”. May be fixed so that only the predicted probability is referred to.
  • an image decoding apparatus that performs a decoding process of a binary shape signal with reference to only a prediction probability instead of the above-described predicted pixel value
  • an image decoding device that performs a decoding process of a multi-level image signal
  • FIG. 10 is a block diagram showing a configuration of an image decoding apparatus according to Embodiment 6 of the present invention.
  • the same reference numerals as those in FIG. 8 denote the same parts as those in the image decoding apparatus 105 according to the fifth embodiment.
  • the image decoding apparatus 106 includes a prediction probability generator 22 instead of the prediction value generator 14 in the image decoding apparatus 105 according to the fifth embodiment.
  • a prediction probability generator 22 instead of the prediction value generator 14 in the image decoding apparatus 105 according to the fifth embodiment.
  • the prediction probability generator 22 determines from the pixel values of the reference pixels ⁇ '′ to ⁇ 9 ′ for the decoded pixel ⁇ ⁇ ′ that the pixel value of the decoded pixel ⁇ ⁇ ′ matches the predicted pixel value. The probability is obtained as the prediction probability S k, and the prediction probability S k for the decrypted pixel P x ′ is output to the decryptor 26 b.
  • the decoder 26b is configured to perform a decoding process on the image coded signal Cs encoded by the image encoding device 102 of the second embodiment.
  • the image decoding apparatus 105 In the decoding processing of the binary shape signal by the image decoding apparatus 105 having such a configuration, even the reference pixels P 0 ′ to ⁇ 9 ′ corresponding to the decoded pixel P x ′ correspond to the undecoded pixels. , The pixel value is generated from the pixel value of the decoded pixel in the same manner as the image decoding apparatus 105 of the fifth embodiment, and the pixel values corresponding to all the reference pixels PC) to ⁇ 9 ′ are predicted. Stored in the probability generator 22.
  • the prediction probability generator 22 obtains a prediction probability S k for the pixel to be decoded ⁇ ⁇ ′ based on the pixel values of the reference pixels ⁇ ⁇ ⁇ to 9 '.
  • the prediction probability S k is output from the prediction probability generator 22 to the decoder 26b
  • the decoder 26b presets the pixel value of the pixel P x ′ to be decoded.
  • a decoding process according to the above-mentioned prediction probability S k is performed on the difference value from the prediction pixel value.
  • the image decoding apparatus 106 that decodes an image coded signal obtained by coding a binary image signal, it is difficult to generate a predicted value adjacent to a block boundary. It is possible to realize a decoding process corresponding to an encoding process in which pixels can be encoded while suppressing deterioration of the encoding efficiency.
  • FIG. 11 is a professional diagram showing the configuration of an image decoding apparatus according to Embodiment 7 of the present invention.
  • FIG. 8 the same reference numerals as those in FIG. 8 denote the same parts as those in the image coding apparatus 105 of the fifth embodiment.
  • the image decoding device 107 performs the encoding process on the multi-valued image signal as described above, and is encoded by the image encoding device 103 according to the third embodiment. In this configuration, the decoded image coded signal Cs is decoded.
  • the prediction probability generator 170 of the image decoding device 107 has a configuration in which the prediction probability generator 22 is provided in addition to the prediction value generator 14 in the image encoding device 105 of the fifth embodiment. Has become.
  • the prediction probability generator 22 receives the output S out of the selection switch 12 and sequentially stores the pixel values of the reference pixels P 0 ′ to P 9 ′ with respect to the decrypted pixel P, similarly to the prediction value generator 14.
  • the decoder 26c performs a decoding process corresponding to the encoding process performed by the encoder 16c according to the third embodiment.
  • the decoder 26c includes the prediction value generator Based on the predicted pixel signal S from 14 and the predicted probability S k from the predicted probability generator 22, the pixel value of the decoded pixel ⁇ ⁇ ′ is decrypted.
  • the pixel value is generated from the pixel value of the decoded pixel in the same manner as in the image decoding device 105 of the fifth embodiment, and the pixels corresponding to all the reference pixels ⁇ ⁇ ′ to ⁇ 9 ′
  • the values are stored in the predicted probability generator 22 and the predicted value generator 14.
  • the predicted value generator 14 generates the reference pixel in the same manner as in the fifth embodiment.
  • the predicted pixel value of the pixel to be decoded ⁇ ⁇ ′ is obtained from P 0 ′ to P 9. Further, the prediction probability generator 22 obtains a prediction probability for the decoded pixel ⁇ ′ based on the pixel values of the reference pixels P CT to ⁇ 9 ′.
  • the decoder 26c When the predicted pixel value S ⁇ and the predicted probability S k are output to the decoder 26c and the local decoder 24, the decoder 26c outputs the pixel value and the pixel value of the pixel P x ′ to be decoded.
  • the decoding processing according to the above-described prediction probability S k is performed on the added value with the predicted pixel value from the predicted value generator 14.
  • the reproduced pixel value of the pixel to be decoded P X ′ is stored in the main memory 4.
  • the seventh embodiment in the image decoding apparatus 107 that decodes an image coded signal obtained by coding a multi-valued image signal, it is difficult to generate a predicted value adjacent to a block boundary. It is possible to realize a decoding process corresponding to an encoding process in which pixels can be encoded while suppressing deterioration in encoding efficiency.
  • the image coded signals that are losslessly coded by the image decoding devices 101, 102, and 103 of the first, second, and third embodiments are used as the image decoding devices.
  • the decoder 26a by configuring the decoder 26a to perform a decoding process corresponding to lossy coding, the image decoding in the fifth, sixth, and seventh embodiments is performed.
  • the devices 105, 106, and 107 can correspond to the image coding devices 104, 10a, and 104b of the fourth embodiment and the first and second modifications.
  • FIG. 12 (a) is a diagram for explaining a case where the encoding or decoding processing of the above embodiment is performed by a computer system using a floppy disk storing the above encoding or decoding program.
  • FIG. 12 (a) is a diagram for explaining a case where the encoding or decoding processing of the above embodiment is performed by a computer system using a floppy disk storing the above encoding or decoding program.
  • Fig. 12 (b) shows the appearance, cross-sectional structure, and floppy disk of the floppy disk as viewed from the front
  • Fig. 12 (a) shows the floppy disk, which is the main body of the recording medium.
  • An example of the physical format of a P.D. disk is shown.
  • the floppy disk FD is built in the case F.
  • On the surface of the disk a plurality of tracks Tr are formed concentrically from the outer circumference toward the inner circumference, and each track has 16 sectors S e in the angular direction. Is divided into Therefore, in the floppy disk storing the program, data as the program is recorded in an area allocated on the floppy disk FD.
  • FIG. 12 (c) shows a configuration for recording and reproducing the above program on a floppy disk FD.
  • the above program is recorded on the floppy disk FD
  • data as the above program is written from the computer system Cs via the floppy disk drive.
  • the encoding or decoding device is constructed in a computer system by a program in a floppy disk
  • the program is read from the floppy disk by a floppy disk drive and transferred to the computer system.
  • the description has been made using a floppy disk as a data recording medium, but the same can be done using an optical disk.
  • the recording medium is not limited to this, and the present invention can be similarly implemented as long as the program can be recorded thereon, such as an IC card or a ROM cassette.
  • the codeword (coding table) may be changed according to the prediction probability.
  • the probability table corresponding to the arithmetic code is updated based on the predicted probability, so that a simple configuration can be used to provide a modification of the second, third, and fourth embodiments. It is possible to realize the image coding device of the first or second embodiment or the image decoding device of the fifth or sixth embodiment. In this case, the practical effect is very large.
  • an uncoded pixel among the peripheral pixels includes The pseudo pixel value is generated using the pixel value of the coded pixel among the pixels, but the position of the coded pixel in the block is calculated.
  • a group of peripheral pixels referred to when generating the predicted pixel value those having different arrangements around the pixel to be coded may be used.
  • the pixel to be coded Px when the pixel to be coded Px is located at the boundary of the block, its surrounding pixels P0, P1, P3, P4, P5 , P8, and P9 as reference pixels, and if the pixel to be coded PX is located outside the block boundary, its surrounding pixels PO, P1, P2, P3, P4, P5, All of P6, P7, P8, and P9 are set as reference pixels. Then, when encoding the pixel to be encoded PX, a codeword composed of only the peripheral pixels PO, P1, P3, P4, P5, P8, and P9 and the peripheral pixel P0 , P1, P2, P3, P4, P5, P6, P7, P8, and P9.
  • the encoder is configured to have a plurality of codewords corresponding to the position of the pixel to be encoded PX in the block, and to switch the codeword according to the position of the pixel to be encoded PX. .
  • the decoder also has a configuration having a plurality of codewords corresponding to the position of the pixel to be decoded Px ′ shown in FIG. 17 in the block, and the codeword corresponding to the position of the pixel to be decoded P described above. It is needless to say that the same effect as that of the image decoding apparatus according to the above-described embodiment can be obtained by adopting a configuration that switches Industrial applicability
  • the image encoding device and the image encoding method, the image decoding device and the image decoding method, and the data recording medium according to the present invention can improve the encoding efficiency in the compression processing of the image signal.
  • This is extremely useful as a means for realizing image encoding and image decoding in a system that transmits and stores image signals.
  • it can compress and compress moving images conforming to standards such as MPEG4. Suitable for elongation processing.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Compression Of Band Width Or Redundancy In Fax (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Image Processing (AREA)

Description

明 細 書 画像符号化装置及び画像符号化方法、 画像復号化装置及び画像複号化方法、 並びにデータ記録媒体 技術分野
本発明は、 画像符号化装置及び画像符号化方法、 画像復号化装置及び画像 復号化方法、 並びにデータ記録媒体に関し、 特に画像信号の記録あるいは伝 送を、 より少ないビット数でもって画質を損なうことなく行うための符号化 処理及び復号化処理、 並びにこれらの符号化処理あるいは複号化処理を実現 するためのプログラムを格納した記録媒体に関するものである。 背景技術
従来の画像符号化処理は、 M P E G 2準拠の符号化方法により代表される ブロック単位で行われる符号化処理と、 差分パルス符号変調 (D P C M : Differential Pulse Code Modulation) 等の画素単位で行われる符号化処理 の 2つに大別される。
プロック単位の符号化処理は、 1画像表示領域を複数のプロックに区分し、 入力される画像信号 (以下、 画像入力信号ともいう。 ) の符号化処理を各ブ ロック毎に行う手法である。 ここで、 上記 1画像表示領域は、 M P E G 2に 準拠した符号化処理では 1つの表示画面に相当するものであり、 また、 M P E G 4に準拠した符号化処理では、 1つの表示画面上の個々の物体 (ォブジ ェク ト) に対応した形状及び大きさを有する表示領域に相当するものである。 また、 上記各プロックは 1画像表示領域内における所定数の画素からなる表 示領域であり、 該ブロックの形状としては処理が容易な矩形形状がよく使用 される。
このようにプロック単位で画像入力信号の符号化処理を行う符号化方法で は、 1画像表示領域に対応する画像入力信号の符号化処理が、 各ブロック毎 に完結することとなる。 このため、 上記画像入力信号に符号化処理を施して 得られる画像符号化信号を伝送する際に伝送誤りが発生しても、 誤りの影響 をブロック単位で収束させることが可能であるという長所がある。
その反面、 上記プロック単位の符号化方法では以下のような欠点がある。 まず、 ブロック単位の符号化方法では、 画像入力信号の符号化処理がプロ ック毎に完結するため、 ブロック間の画素相関, つまり異なるブロック間に 存在する画素値の相関性を符号化処理の際に利用することが困難である。 また、 画像信号の予測符号化方法では、 符号化の対象となる被符号化画素 の画素値 (被符号化画素値) を、 これより先に符号化した複数の符号化済画 素の画素値 (符号化済画素値) を参照して予測し、 該予測した画素値を用い て被符号化画素値を適応的に符号化する処理を行っている。 ところが、 この ような予測符号化方法において符号化処理をプロック単位で行うと、 被符号 化画素値の符号化の際に参照される符号化済画素値がプロック内の画素に限 定され、 該参照される符号化済画素値の数が少なくなる。 このため、 被符号 化画素の予測値の確からしさが小さくなり、 符号化効率があまり高くならな い。
一方、 画素単位の符号化方法は、 入力される画像信号を画素毎に符号化す る手法であり、 この符号化方法では、 画像入力信号の符号化処理を画素単位 で変更することも可能である。このため、この符号化方法において、例えば、 画 像入力信号の特性に合わせて符号語が画素毎に自動的に更新される適応算術 符号化等のユニバーサル符号化処理を行うようにすれば、 どのような特¾^の 画像信号も非常に高い符号化効率でもって符号化することができる。
ところが、 上記ユニバーサル符号化処理を行う画素単位の符号化方法によ り得られた画像符号化信号には、 復号化側にて、 符号化側と全く同様に符号 語を更新する復号化処理が施されるため、 上記画像符号化信号を伝送する際 に伝送誤りが発生すれば、 復号化側では、 該伝送誤りの影響により、 画像符 号化信号の復号化処理を正しく行うことができない状態が長く続く という短 所がある。
ところで、 上記プロック単位の符号化方法と画素単位の符号化方法は組み 合わせることも可能であり、 これらを組み合わせた符号化方法 (以下、 説明 の都合上この符号化方法を組合せ符号化方法という。 ) では、 画素毎に適応 的に符号語を切り替え、 かつ伝送誤りの影響をプロック単位で収束すること が可能となり、 符号化効率の高い適応算術符号化等の符号化処理を伝送誤り の影響を小さく抑えつつ行うことができる。
以下、 この組合せ符号化方法について説明する。
第 1 3 (a) 図は 1フレーム画面を複数の矩形形状のブロックに区分した状 態を示しており、 第 1 3 (b) 図はブロック内, 特に符号化対象となっている 被符号化ブロック及びその周辺のブロック内での画素の配列を示している。 なお、 各画素は、 1フレーム画面内の各水平走査線に沿ってマトリクス状に 配列されていることは言うまでもない。
図中、 F Gは 1フレームに対応する画面、 B 1はすでに画像信号に対する 符号化処理が施された符号化済ブロック、 B Xは符号化処理の対象となって いる被符号化プロック、 B 0は未だ画像信号に対する符号化処理が施されて いない未符号化ブロックである。 ただし、 上記各ブロックを区別しない場合 はブロック Bと記載する。 B L u, B L s, B L h , 8 111は1フレーム画 面上での被符号化ブロックの上, 下, 左, 右の境界である。 また、 実線の丸 印はその画素値に対する符号化処理が既に施された符号化済画素を示し、 点 線の丸印はその画素値に対する符号化処理が未だ施されていない未符号化画 素を示している。 また、 上記各ブロック Bは、 上記 1フレーム画面 F G上に おける、 4 X 4画素からなる画像表示領域である。
第 1 4図は、 符号化処理の対象となる被符号化画素 P Xと、 その周辺に位 置する周辺画素 P 0〜P 9との位置関係を示しており、 この周辺画素 P 0〜 P 9はその画素値が、 被符号化画素 P Xの画素値を予測する際に参照される ものであり、 以下参照画素 P 0〜P 9という。
ここで、 上記参照画素 P 8, P 9は上記被符号化画素 P Xと同一の水平走 查線上に位置する画素であり、 該参照画素 P 9, P 8はそれぞれ、 上記被符 号化画素 P xの 1画素分, 2画素分だけ手前に位置している。 また、 上記参 照画素 P 5 , P 1の、 1フレーム画面 F G上での水平方向における位置は、 上記被符号化画素 P Xと同一であり、 上記参照画素 P 5, P 1はそれぞれ、 上記被符号化画素 P Xの 1画素分, 2画素分だけ上の水平走査線上に位置し ている。 さらに、 上記参照画素 P 3, P 4, P 6, P 7は上記参照画素 P 5 と同一の水平走査線上に位置する画素であり、 該参照画素 P 4, P 3はそれ ぞれ上記被符号化画素 P Xの 1画素分, 2画素分だけ手前に位置し、 該参照 画素 P 6, P 7はそれぞれ上記被符号化画素 P Xの 1画素分, 2画素分だけ 後ろに位置している。 さらに、 上記参照画素 P 0, P 2は上記参照画素 P 1 と同一の水平走査線上に位置する画素であり、 該参照画素 P 0は上記参照画 素 P 1の 1画素分だけ手前に、 該参照画素 P 2は上記参照画素 P 1の 1画素 分だけ後ろに位置している。
上記組合せ符号化方法では、 まず、 第 1 3 (a) 図、 第 13図(b) に示すよ うに、 1フレーム画面 FGに対応する画像信号が、 該 1フレーム画面を構成 する複数のプロック Bの各々に対応するよう分割され、 該分割された画像信 号の符号化処理が各ブロック毎に行われる。
そして、 このブロック単位の符号化処理は、 ブロック B内における横方向 画素列に沿つて順次左側から右側へ向かって各画素の画素値を符号化する水 平処理を、 上記各プロック内の全ての横方向画素列について最上列から順次 最下列まで行うことにより完結する。
また、 この符号化処理では、 第 14図に示すように、 被符号化画素 P xの 画素値を、 それぞれ被符号化画素の周辺に位置する参照画素 P 0〜P 9の画 素値から適応的に予測し、 予測により得られた予測値に応じて、 被符号化画 素の符号化処理に用いる符号語を適応的に選択している。
これにより、 上記組合せ符号化方法では、 伝送誤りの復号化側での影響を ブロック単位で収束可能とし、 しかも、 単純なブロック単位の符号化処理に 比べて符号化効率を向上することができる。
一方、 第 16 (a) 図, 第 16 (b) 図及び第 1 7図は、 それぞれ上記組合せ 符号化方法に対応する組合せ複号化方法を説明するための図であり、 図中、 B' は 1フレーム画面における各ブロックであり、 B x' は被複号化ブロッ ク、 B 1' は複号化済ブロック、 B 0' は未復号化ブロック、 B Lu' , B L s ' , B Lh' , B Lm' は、 被復号化ブロック B x' の上, 下, 左, 右 側のブロック境界であり、 Ρ Ο' 〜Ρ 9' は被復号化画素 Ρ χ' に対応する 参照画素である。 ここで、 被複号化画素 Ρ χ' に対する参照画素 Ρ Ο' 〜Ρ 9' の配置は、 第 1 3 (a) 図, 第 1 3 (b) 図, 第 14図に示す符号化処理の ものと全く同一となっている。
上記組合せ複号化方法では、 まず、 第 1 6 (a) 図、 第 1 6図(b) に示すよ うに、 1フレーム画面 FG' に対応する画像信号が、 該 1フレーム画面を構 成する複数のブロック B' の各々に対応するよう分割され、 該分割された画 像信号の復号化処理が各プロック毎に行われる。
そして、 このブロック単位の複号化処理は、 ブロック B' 内における横方 向画素列に沿つて順次左側から右側へ向かって各画素の画素値を復号化する 水平処理を、 上記各プロック内の全ての横方向画素列について最上列から順 次最下列まで行うことにより完結する。
また、 この復号化処理では、 第 1 7図に示すように、 被復号化画素 Ρ χ' の画素値を、 それぞれ被複号化画素の周辺に位置する参照画素 Ρ 0' 〜Ρ 9' の画素値から適応的に予測し、 予測により得られた予測値に応じて、 被復号 化画素 Ρ χ' の復号化処理に用いる符号語を適応的に選択している。
しかしながら、 上記プロック単位の符号化処理と画素単位の符号化処理の 両者を組合わた組合せ符号化方法では、 以下のような問題があった。
この組合せ符号化方法では、符号化処理はプロック単位で進められるため、 第 1 5図に示すように被符号化画素 Ρ Xが被符号化プロック Β Xの右側境界 BLmに隣接して位置する場合、該被符号化画素 P Xに対する参照画素 P 2, P 6, P 7が未符号化画素となる。
この場合、 上記未符号化画素 P 2, P 6, P 7の画素値を参照して被符号 化画素 P Xの画素値を予測し、 この予測値を用いて被符号化画素 P Xの画素 値を符号化すると、 復号化側でこの被符号化画素 P Xに対応する画像符号化 信号を正しく複号化することが不可能となる。
つまり、 被符号化画素 P Xの画素値をその予測値を用いて符号化して得ら れる画像符号化信号を復号化側で正しく復号化するには、 復号化処理の際に 用いる被複号化画素 Ρ χ' の予測値を、 符号化処理の際に用いる、 該被復号 化画素 P に対応する被符号化画素 Ρ χの予測値と一致させる必要がある。 これは、 言い換えると、 符号化側では、 被符号化画素 Ρ Xの予測値を生成す る際に参照される参照画素値が、 被符号化画素 Ρ Xに対応する被復号化画素 Ρ χ' の予測値を生成する際に参照される参照画素値と完全に一致するよう にしておく必要があるということである。
このため、 例えば、 第 1 5図に示すように、 被符号化画素 Ρ χに対する符 号化処理を行う際に、 被符号化画素 Ρ Xに対する参照画素 Ρ 0〜Ρ 9のうち の未符号化画素 Ρ 2, Ρ 6, Ρ 7の画素値を参照して、 被符号化画素 Ρ Xの 予測値を生成した場合、 複号化側では、 第 1 8図に示すように、 被復号化画 素 Ρ χ' に対する復号化処理を行う際、 被復号化画素 Ρ χ' に対応する参照 画素 Ρ Ο' 〜Ρ 9' の画素値を参照して被複号化画素 Ρ χ' の予測値を生成 することとなるが、 復号化側では、 上記参照画素 Ρ Ο' 〜Ρ 9' のうちの未 復号化画素 Ρ 2' , P 6' , Ρ 1' については画素値が得られず、 上記被符 号化画素 Ρ Xに対応する被復号化画素 Ρ X ' の画素値を復号化することがで きない。
そこで、 従来の組合せ符号化方法では、 上記のように被符号化画素 Ρ Xに 対応する参照画素 Ρ 0〜Ρ 9に未符号化画素が含まれる場合に複号化処理が 困難となるという問題を回避するため、 上記未符号化画素については、 その 画素値を予め設定された固定値 (例えば 0) と見なして、 被符号化画素 Ρ χ の予測値を生成し、 この予測値を用いて被符号化画素 Ρ Xに対する符号化処 理を行うという対応策を講じている。
このような対応策を講じた組合せ符号化方法では、 複号化側では、 ブロッ ク内のすベての画素についてその予測値を用いた複号化処理を正しく行うこ とが可能になるが、 未符号化画素となる参照画素の画素値を一律に固定値と 置き換えるため、 未符号化画素と符号化済画素の間での画素値の相関関係が 損なわれることとなり、 この結果、 被符号化画素の予測効率, つまり被符号 化画素に対する予測値の確からしさが劣化するという問題が生ずる。
本発明は上記のような課題を解決するためになされたもので、 画素単位で の適応的な符号化処理とプロック単位での符号化処理とを、 未符号化画素と 被符号化画素との間での画素値の相関を損なうことなく、 しかも画像符号化 信号の復号化が困難となるのを回避しつつ組み合わせることができる画像処 理装置及び画像処理方法、 並びに該画像処理方法を実現するための画像処理 プログラムを記憶したデータ記憶媒体を得ることを目的とする。
また、 本発明は、 被符号化画素の予測効率の劣化を招くことなく符号化さ れた画像符号化信号の複号化処理を正しく行うことができる画像処理装置及 び画像処理方法、 並ぴに該画像処理方法を実現するための画像処理プロダラ ムを記憶したデータ記憶媒体を提供することを目的とする。 発明の開示
本発明 (請求の範囲第 1項) に係る画像処理装置は、 画像信号を構成する 画素値を、 被符号化画素の周辺に位置する複数の周辺画素の画素値に基づい て順次符号化する画像符号化装置であって、 1画像表示領域に対応する複数 の画素値からなる画像信号を、 所定数の画素からなるプロックに対応するよ うブロック化し、 各ブロックの画像信号を構成する所定数の画素値をブロッ ク単位で出力するプロック化手段と、 上記周辺画素のうちの未符号化画素の 画素値を、 上記周辺画素のうちの符号化済画素の画素値から所定の規則に基 づいて得られる擬似画素値と置換する画素値置換手段と、 上記各プロックに 対応する複数の画素値からなる画像信号を受け、 該各画素値を上記符号化済 画素の画素値及び未符号化画素の擬似画素値に基づいて順次符号化する符号 化処理を各プロック毎に行って画像符号化信号を出力する符号化手段とを備 えたものである。
このような構成の画像符号化装置によれば、 被符号化画素に対応する複数 の周辺画素のうちの未符号化画素の画素値を、 上記複数の周辺画素のうちの 符号化済画素の画素値に基づいて得られる擬似画素値に置き換える画素値置 換手段を備えたので、 1画像表示領域に対応する画像信号をプロック単位で 符号化する処理を、 被符号化画素周辺の周辺画素の画素値を画素毎に参照し て行う際、 被符号化画素がブロックの境界に隣接して位置し、 参照される複 数の周辺画素に未符号化画素が含まれる場合であっても、 未符号化画素の画 素値として、 他の周辺画素の画素値との間で相関のとれた擬似画素値を参照 することができる。 この擬似画素値は被符号化画素周辺の符号化済画素の画 素値から得られるものであるので、 復号化側においても、 被復号化画素の復 号化の際に参照される未復号化画素については、 その画素値の代わりに、 復 号化済画素の画素値から得られる擬似画素値を参照することができる。
このため、 画素単位での適応的な符号化処理とプロック単位での符号化処 理とを、 未符号化画素と被符号化画素との間での画素値の相関を損なうこと なく、 しかも複号化側にて符号化信号の復号化が困難となるのを回避しつつ 組み合わせることができる。
これにより伝送誤りの影響をブロック単位で収束可能とし、 かつ符号化効 率を単純なプロック単位の符号化処理に比べてより高いものとでき、 また、 被符号化画素の予測効率の劣化を招くことなく符号化した符号化信号の復号 化処理を正しく行うことができるという効果が得られる。
この発明 (請求の範囲第 2項) は、 請求の範囲第 1項記載の画像符号化装 置において、 上記画素値置換手段を、 上記未符号化画素からの空間距離が最 短である符号化済画素の画素値を該未符号化画素の擬似画素値とする構成と したものである。
このような構成の画像符号化装置によれば、 未符号化画素の擬似画素値と して、 最も未符号化画素に近い符号化済画素の画素値が用いられるので、 未 符号化画素の擬似画素値を、 周辺画素の画素値との相関の強いものとできる。 この発明 (請求の範囲第 3項) は、 請求の範囲第 1項記載の画像符号化装 置において、 上記画素値置換手段を、 上記未符号化画素からの空間距離が最 短である、 該未符号化画素と同じ水平走査線上に位置する符号化済画素の画 素値を、 該未符号化画素の擬似画素値とする構成としたものである。
このような構成の画像符号化装置によれば、 未符号化画素の擬似画素値と して、 未符号化画素と同一水平走査線上で上記未符号化画素から最も近い符 号化済画素の画素値を用いるので、 符号化済画素の画素値を所定時間保持す るといった簡単な方法により、 未符号化画素の擬似画素値を得ることができ る。 この発明 (請求の範囲第 4項) は、 請求の範囲第 1項記載の画像符号化装 置において、 上記符号化手段を、 上記符号化済画素の画素値及び上記未符号 化画素の擬似画素値に基づいて被符号化画素に対する予測画素値を生成する 予測値生成器と、 上記被符号化画素の画素値と上記被符号化画素の予測画素 値との差分値を符号化し、 該符号化した差分値をブロック単位で画像符号化 信号として出力する符号化器とを有する構成としたものである。
このような構成の画像符号化装置によれば、 上記被符号化画素の画素値と その予測画素値との差分値を符号化するので、 被符号化画素の画素値の符号 化に必要となる符号量を低減して符号化効率を向上するこができる。
この発明 (請求の範囲第 5項) は、 請求の範囲第 1項または第 4項に記載 の画像符号化装置において、 上記符号化手段を、 上記符号化済画素の画素値 及び上記未符号化画素の擬似画素値に基づいて、 上記被符号化画素の画素値 を符号化するための符号語を選択する構成としたものである。
このような構成の画像符号化装置によれば、 上記符号化済画素の画素値及 び未符号化画素の擬似画素値に基づいて、 被符号化画素の画素値を符号化す るための符号語を選択するようにしたので、 画像信号の符号化処理の際、 画 素毎に符号語を適応的に変更して、 効率のよい符号化処理を実現することが できる。
この発明 (請求の範囲第 6項) は、 請求の範囲第 1項または第 4項に記載 の画像符号化装置であって、 上記符号化手段を、 上記符号化済画素の画素値 及び上記未符号化画素の擬似画素値に基づいて、 上記被符号化画素の画素値 を算術符号化するための符号コードに相当する確率テーブルを選択し、 上記 被符号化画素に対する算術符号化処理を、 該選択した確率テーブルに基づい て行う構成としたものである。
このような構成の画像符号化装置によれば、 上記符号化済画素の画素値及 び未符号化画素の予測画素値に基づいて、 被符号化画素を符号化する算術符 号に相当する確率テーブルを選択して、 被符号化画素に対する画素値の符号 化を行うので、 算術符号化処理における確率テーブルを画素毎に適応的に切 り替えて、 効率のよい符号化処理を行うことができる。 本発明 (請求の範囲第 7項) は、 請求の範囲第 1項記載の画像符号化装置 において、 上記符号化済画素の画素値を複号化して局所復号化画素値を生成 する局所復号化手段を備え、 上記画素値置換手段を、 上記未符号化画素の画 素値を、 上記符号化済画素に対応する局所復号化画素値から所定の規則に基 づいて得られる擬似画素値と置換する構成とし、 上記符号化手段を、 上記符 号化済画素の局所複号化画素値及び上記未符号化画素の擬似画素値に基づい て、 上記被符号化画素の画素値に対して非可逆符号化処理を施す構成とした ものである。
このような構成の画像処理装置によれば、 被符号化画素の画素値に対して 非可逆符号化処理を、 上記符号化済画素の画素値を複号化した複号化画素値 に基づいて行うので、 被符号化画素に対する非可逆符号化処理の際に参照す る周辺画素の画素値を、 被復号化画素に対する複号化処理の際に参照される 周辺画素の画素値と同一にすることができ、 これにより周辺画素の画素値に 基づいた非可逆符号化により得られる画像符号化信号を、 複号化側にて正し く復号化することが可能となる。
この発明 (請求の範囲第 8項) は、 請求の範囲第 7項記載の画像符号化装 置において、 上記符号化手段を、 上記符号化済画素の局所復号化画素値及び 上記未符号化画素の擬似画素値に基づいて被符号化画素に対する予測画素値 を生成する予測値生成器と、 上記被符号化画素の画素値と上記被符号化画素 の予測画素値との差分値を符号化し、 該符号化した差分値をプロック単位で 画像符号化信号として出力する符号化器とを有する構成としたものである。 このような構成の画像符号化装置によれば、 上記符号化済画素の画素値お よび未符号化画素の擬似画素値から被符号化画素の予測画素値を生成し、 上 記被符号化画素の画素値とその予測画素値との差分値を符号化するので、 被 符号化画素の画素値の符号化に必要となる符号量を低減して符号化効率を向 上するこができる。
この発明 (請求の範囲第 9項) は、 請求の範囲第 7項または第 8項に記載 の画像符号化装置において、 上記符号化手段を、 上記符号化済画素の局所復 号化画素値及び上記未符号化画素の擬似画素値に基づいて、 上記被符号化画 素の画素値を符号化するための符号語を選択する構成としたものである。 このような構成の画像符号化装置によれば、 上記符号化済画素の画素値及 び未符号化画素の擬似画素値に基づいて、 該被符号化画素の画素値を符号化 するための符号語を選択するので、 画像信号の符号化処理の際、 画素毎に符 号語を適応的に変更して、 効率のよい符号化処理を実現することができる。 この発明 (請求の範囲第 1 0項) は、 請求の範囲第 7項または第 8項に記 載の画像符号化装置において、 上記符号化手段を、 上記符号化済画素の画素 値及び上記未符号化画素の擬似画素値に基づいて、 上記被符号化画素の画素 値を算術符号化するための符号コードに相当する確率テーブルを選択し、 上 記被符号化画素に対する算術符号化処理を、 該選択した確率テーブルに基づ いて行う構成としたものである。
このような構成の画像符号化装置によれば、 上記符号化済画素の画素値及 び未符号化画素の擬似画素値に基づいて、 被符号化画素を符号化する算術符 号に相当する確率テーブルを選択して、 被符号化画素に対する画素値の符号 化を行うので、 算術符号化処理における確率テーブルを画素毎に適応的に切 り替えて、 効率のよい符号化処理を行うことができる。
この発明 (請求の範囲第 1 1項) に係る画像複号化装置は、 画像信号を構 成する画素値を、 被符号化画素の周辺に位置する複数の周辺画素の画素値に 基づいて順次符号化する処理を、 所定数の画素からなるブロック毎に行って 得られた画像符号化信号を、 該ブロック毎に復号化する画像複号化装置であ つて、 被複号化画素の周辺に位置する複数の周辺画素のうちの未復号化画素 の画素値を、 該複数の周辺画素のうちの復号化済画素の画素値から所定の規 則に基づいて得られる擬似画素値と置換する画素値置換手段と、 上記各プロ ックに対応する複数の画素値からなる画像符号化信号を受け、 該各画素値を 上記復号化済画素の画素値及び未復号化画素の擬似画素値に基づいて順次復 号化する復号化処理を、 上記ブロック単位で行って、 各ブロックに対応した 画像復号化信号を出力する複号化手段と、 上記各プロックに対応した画像復 号化信号を統合して、 走査線構造の画像複号化信号に変換する逆プロック化 手段を備え、 該走査線構造の画像複号化信号を 1画像表示画面に対応する画 像再生信号として出力するものである。
このような構成の画像複号化装置によれば、 被復号化画素周辺に位置する 複数の周辺画素のうちの未復号化画素の画素値を、 上記複数の周辺画素のう ちの復号化済画素の画素値に基づいて得られる擬似画素値に置き換える画素 値置換手段を備えたので、 1画像表示領域に対応する画像符号化信号をプロ ック単位で復号化する処理を、 被複号化画素周辺の周辺画素の画素値を画素 毎に参照して行う際、 被復号化画素がブロックの境界に隣接して位置し、 参 照される複数の周辺画素に未複号化画素が含まれる場合であっても、 未復号 化画素の画素値として、 他の周辺画素の画素値との間で相関のとれた擬似画 素値を参照することができる。
このため、 画素単位での適応的な複号化処理とプロック単位での複号化処 理とを、 未復号化画素と被復号化画素との間での画素値の相関を損なうこと なく組み合わせた複号化方法を実現できる。 これにより、 画素単位での適応 的な符号化処理とプロック単位での符号化処理とを組み合わせた符号化方法 により符号化した画像符号化信号を、 正しく復号化することができる効果が ある。
この発明 (請求の範囲第 1 2項) は、 請求の範囲第 1 1項記載の画像復号 化装置において、 上記画素値置換手段を、 上記未復号化画素からの空間距離 が最短である複号化済画素の画素値を該未復号化画素の擬似画素値とするよ う構成したものである。
このような構成の画像復号化装置によれば、 未複号化画素の擬似画素値と して、 最も未複号化画素の近い復号化済画素の画素値が用いられるので、 未 復号化画素の擬似画素値を、 周辺画素の画素値との相関の強いものとできる。 この発明 (請求の範囲第 1 3項) は、 請求の範囲第 1 1項記載の画像復号 化装置において、 上記画素値置換手段を、 上記未複号化画素からの空間距離 が最短である、 該未復号化画素と同じ水平走査線上に位置する複号化済画素 の画素値を、 該未復号化画素の擬似画素値とするよう構成したものである。 このような構成の画像複号化装置によれば、 未複号化画素の擬似画素値と して、 未複号化画素と同一水平走査線上で上記未復号化画素から最も近い復 号化済画素の画素値を用いるので、 復号化済画素の画素値を所定時間保持す るといった簡単な方法により、 未復号化画素の擬似画素値を得ることができ る。
この発明 (請求の範囲第 1 4項) は、 請求の範囲第 1 1項記載の画像復号 化装置において、 上記復号化手段を、 上記複号化済画素の画素値及び未復号 化画素の擬似画素値に基づいて被復号化画素に対する予測画素値を生成する 予測値生成器と、 上記被復号化画素の画素値を複号化して復号化画素値を生 成し、 該復号化画素値に該被復号化画素の予測画素値を加算する復号化器と を有する構成としたものである。
このような構成の画像復号化装置によれば、 上記複号化済画素の画素値及 び未複号化画素の擬似画素値から被復号化画素の予測画素値を生成し、 復号 化された被復号化画素の画素値に該予測画素値を加えるので、 被符号化画素 の画素値とその予測画素値との差分値に対する符号化処理により得られた画 像符号化信号を、 正しく複号化することができる。
この発明 (請求の範囲第 1 5項) は、 請求の範囲第 1 1項又は第 1 4項に 記載の画像復号化装置において、 上記復号化手段を、 上記復号化済画素の画 素値および上記未複号化画素の擬似画素値に基づいて、 上記被複号化画素の 画素値を複号化するための符号語を選択する構成としたものである。
このような構成の画像復号化装置によれば、 上記複号化済画素の画素値お よび未複号化画素の擬似画素値に基づいて、 該被復号化画素の画素値を復号 化するための符号語を選択するようにしたので、 画素毎に符号語を適応的に 変更して効率のよい符号化処理を施した画像符号化信号を、 正しく復号化す ることが可能となる。
この発明 (請求の範囲第 1 6項) は、 請求の範囲第 1 1項又は第 1 4項に 記載の画像複号化装置において、 上記複号化手段を、 上記複号化済画素の画 素値及び未複号化画素の擬似画素値に基づいて、 上記被複号化画素の画素値 を算術復号化するための符号コ一ドに相当する確率テーブルを選択し、 上記 被復号化画素に対する算術復号化処理を、 該選択した確率テーブルに基づい て行うよう構成したものである。 このような構成の画像復号化装置によれば、 上記複号化済画素の画素値及 び未復号化画素の擬似画素値に基づいて、 被複号化画素を復号化する算術符 号に相当する確率テーブルを選択して、 被復号化画素に対する画素値の復号 化を行うので、 確率テーブルを画素毎に適応的に切り替えて算術符号化処理 を施した画像符号化信号を、 正しく復号化することができる。
この発明 (請求の範囲第 1 7項) に係る画像符号化方法は、 画像信号を構 成する画素値を、 被符号化画素の周辺に位置する複数の周辺画素の画素値に 基づいて順次符号化する符号化処理を、 所定数の画素からなるプロック毎に 行う画像符号化方法であって、 上記複数の周辺画素のうちの符号化済画素に ついては、 その画素値を参照画素値として設定し、 かつ上記複数の周辺画素 のうちの未符号化画素については、 該周辺画素のうちの符号化済画素の画素 値から所定の規則に基づいて得られる擬似画素値を参照画素値として設定し、 上記被符号化画素に対する複数の周辺画素に対して設定された参照画素値に 基づいて、 該被符号化画素の画素値を符号化して、 上記画像信号に対する画 像符号化信号を生成するものである。
このような構成の画像符号化方法によれば、 画像信号を 1画像表示領域上 の各プロックに対応するよぅブロックイ匕し、 該ブロックにおける被符号化画 素の画素値を周辺画素の画素値を参照して符号化し、 この際、 周辺画素が符 号化済画素であればその画素値を参照し、 該周辺画素が未符号化画素であれ ば、 その画素値に代えて符号化済画素の画素値から得られる擬似画素値を参 照するようにしたので、 上記符号化処理の際、 被符号化画素がプロックの境 界に隣接して位置し、 参照される複数の周辺画素に未符号化画素が含まれる 場合であっても、 未符号化画素の画素値として、 他の周辺画素の画素値との 間で相関のとれた擬似画素値を参照することができる。 この擬似画素値は被 符号化画素周辺の符号化済画素の画素値から得られるものであるので、 復号 化側においても、 被復号化画素の複号化の際に参照される未復号化画素につ いては、 その画素値に代わりに、 複号化済画素の画素値から得られる擬似画 素値を参照することができる。
このため、 画素単位での適応的な符号化処理とプロック単位での符号化処 理とを、 未符号化画素と被符号化画素との間での画素値の相関を損なうこと なく、 しかも複号化側にて符号化信号の復号化が困難となるのを回避しつつ 組み合わせることができる。
これにより伝送誤りの影響をプロック単位で収束可能とし、 かつ符号化効 率を単純なブロック単位の符号化処理に比べてより高いものとでき、 また、 被符号化画素の予測効率の劣化がなく符号化信号の復号化処理を正しく行う ことができるという効果が得られる。
この発明 (請求の範囲第 1 8項) に係る画像複号化方法は、 画像信号を構 成する画素値を、 被符号化画素の周辺に位置する複数の周辺画素の画素値に 基づいて順次符号化する処理を、 所定数の画素からなるブロック毎に行って 得られた画像符号化信号を、 該ブロック毎に復号化する画像複号化方法であ つて、 上記複数の周辺画素のうちの復号化済画素については、 その画素値を 参照画素値として設定し、 かつ上記複数の周辺画素のうちの未復号化画素に ついては、 該周辺画素のうちの復号化済画素の画素値から所定の規則に基づ いて得られる擬似画素値を参照画素値として設定し、 上記被復号化画素に対 する複数の周辺画素に対して設定された参照画素値に基づいて、 該被復号化 画素の画素値を復号化して、 上記画像符号化信号に対する画像復号化信号を 生成するものである。
このような構成の画像復号化方法によれば、 複数の画素からなるプロック 毎に、 各ブロックにおける被複号化画素の画素値を周辺画素の画素値を参照 して順次復号化する処理を行い、 この際、 周辺画素が複号化済画素であれば その画素値を参照し、 該周辺画素が未復号化画素であれば、 その画素値に代 えて複号化済画素の画素値から得られる擬似画素値を参照するようにしたの で、 上記複号化処理の際、 被復号化画素がブロックの境界に隣接して位置し、 参照される複数の周辺画素に未複号化画素が含まれる場合であっても、 未復 号化画素の画素値として、 他の周辺画素の画素値との間で相関のとれた擬似 画素値を参照することができる。
このため、 画素単位での適応的な複号化処理とプロック単位での複号化処 理とを、 未複号化画素と被複号化画素との間での画素値の相関を損なうこと なく組み合わせた複号化方法を実現できる。 これにより、 画素単位での適応 的な符号化処理とプロック単位での符号化処理とを組み合わせた符号化方法 により符号化した画像符号化信号を、 正しく復号化することができる効果が める。
この発明 (請求の範囲第 1 9項) に係るデータ記憶媒体は、 画像信号の符 号化処理あるいは複号化処理を行うためのプログラムを格納したデータ記憶 媒体であって、 上記プログラムを、 請求の範囲第 1 7項記載の画像符号化方 法または第 1 8項記載の画像復号化方法による画像信号の処理をコンビユー タに行わせるための画像処理プログラムとしたものである。
このようなデータ記憶媒体を用いることにより、 伝送誤り発生時の誤り伝 播をブロック単位で収束可能とするとともに、 単純なブロック単位の符号化 処理に比べて符号化効率を向上することができ、 しかも被符号化画素の予測 効率の劣化がなく符号化信号の復号化処理を正しく行い、 また画素単位での 適応的な符号化処理とブロック単位での符号化処理とを組み合わせた符号化 方法により符号化した画像符号化信号を、 正しく復号化することができる。 図面の簡単な説明
第 1図は、 本発明の実施の形態 1による画像符号化装置の構成を示すプロ ック図である。
第 2図は、 上記画像符号化装置による符号化処理における未符号化画素の 画素値を生成する処理をフローチヤ一トにより示す図である。
第 3図は、 本発明の実施の形態 2による画像符号化装置の構成を示すプロ ック図である。
第 4図は、 本発明の実施の形態 3による画像符号化装置の構成を示すプロ ック図である。
第 5図は、 本発明の実施の形態 4による画像符号化装置の構成を示すプロ ック図である。
第 6図は、 上記実施の形態 4の変形例 1による画像符号化装置の構成を示 すブロック図である。 第 7図は、 上記実施の形態 4の変形例 2による画像符号化装置の構成を示 すブロック図である。
第 8図は、 本発明の実施の形態 5による画像複号化装置の構成を示すプロ ック図である。
第 9図は、 上記画像復号化装置による復号化処理における未復号化画素の 画素値を生成する処理をフローチャートにより示す図である。
第 1 0図は、 本発明の実施の形態 6による画像復号化装置の構成を示すブ 口ック図である。
第 1 1図は、 本発明の実施の形態 7による画像複号化装置の構成を示すブ ロック図である。
第 1 2 (a) 図, 第 1 2 (b) 図, 第 1 2 (c) 図は、 上記実施の形態の画像符 号化装置による符号化処理あるいは画像復号化装置による復号化処理をコン ピュータシステムにより実現するためのプログラムを格納したデータ記憶媒 体を示す図である。
第 1 3 (a) 図、 第 1 3 (b) 図は、 ブロック単位の符号化処理における 1フ レーム画面を複数のブロックに区分する様子を示す模式図である。
第 1 4図は、 画素単位での適応的な符号化処理を説明するための模式図で ある。
第 1 5図は、 上記ブロック単位の符号化処理と画素単位での符号化処理と を組み合わせた組合せ符号化方法における問題点を説明するための模式図で ある。
第 1 6 (a) 図、 第 1 6 (b) 図は、 ブロック単位の複号化処理における 1フ レーム画面を複数のブロックに区分する様子を示す模式図である。
第 1 7図は、 画素単位での適応的な復号化処理を説明するための模式図で ある。
第 1 8図は、 上記ブロック単位の複号化処理と画素単位での複号化処理と を組み合わせた組合せ複号化方法における問題点を説明するための模式図で ある。 発明を実施するための最良の形態
以下、 本発明の実施の形態について、 第 1図から第 1 2図を用いて説明す る。
実施の形態 1 .
第 1図は本発明の実施の形態 1による画像符号化装置 1 0 1の構成を示す ブロック図である。
この画像符号化装置 1 0 1は、 入力された画像信号 I sを、 1画像表示領 域 (1フレーム) を構成する複数のブロックの各々に対応するよう分割する ブロック化器 2と、 該ブロック化器 2の出力 B sを受け、 符号化対象となる 被符号化ブロックにおける、 符号化対象となる被符号化画素の画素値を、 該 被符号化画素の予測画素値 (以下、 単に予測値ともいう。 ) S pを参照して 可逆符号化する符号化器 1 6 aと、 上記予測画素値 S pを生成する予測値生 成部 1 1 0とを有している。
この予測値生成部 1 1 0は、 入力される画像信号を構成する各画素の画素 値を、 例えば 1フレームに相当する数だけ記憶可能な大容量の主メモリ 4と、 該主メモリ 4から出力される画素値 Mを一時的に保持する保持時間の異なる 第 1, 第 2補助メモリ 6 a, 6 bとを有している。 ここで、 上記主メモリ 4 は、 1つの画素値の符号化処理に要する時間の間に、 上記符号化器 1 6 aに て処理の対象となっている被符号化画素 P Xの参照画素 P 0〜P 9 (第 1 4 図参照) に対応する画素値を、 上記記憶した画素値から順次出力する構成と なっている。 また、 上記第 1補助メモリ 6 aは、 主メモリ 4から順次出力さ れる画素値 Mを 1画素分だけ遅延する構成となっており、 上記第 2補助メモ リ 6 bは主メモリ 4から順次出力される画素値 Mを 2画素分だけ遅延する構 成となっている。
また、 上記予測値生成部 1 1 0は、 各フレームの画像入力信号 I sを受け、 画素値の数をカウントするカウンタ 8と、 該カウンタ 8の出力 C o u t及び 外部から供給される各フレームにおける縦横のプロック数の情報 B N s基づ いて、 主メモリ 4から出力されている画素値が、 すでに符号化器 1 6 aにて 符号化処理が施された符号化済画素の画素値であるか、 まだ符号化処理が施 されていない未符号化画素の画素値であるかを判定する符号化/未符号化判 定器 1 0とを有している。 ここで該判定器 1 0は、 未符号化画素については、 これと同一水平走査線上に位置する最も近い符号化済画素までの距離を画素 数でもって計測する処理も行っている。 また、 上記カウンタ 8は、 1フレー ムを構成するすべての画素の画素値が入力された時点でリセットされるよう になっている。
さらに、 上記予測値生成部 1 1 0は、 上記符号化/未符号化判定器 1 0の 出力に基づいて上記主メモリ 4, 第 1, 第 2補助メモリ 6 a, 6 bのいずれ かの出力 M, M a, M bを選択して出力する選択スィッチ 1 2と、 該選択ス イッチ 1 2の出力 S o u tを、 被符号化画素 P Xに対応する予測画素値の生 成に必要な参照画素 P 0〜P 9の画素値として取得し、 該被符号化画素 P X に対応する予測画素値 S pを生成する予測値生成器 1 4とを有している。 そして、 この画像符号化装置 1 0 1では、 上記符号化器 1 6 aは、 プロッ ク化器 2から出力される被符号化画素 P Xの画素値とその予測画素値との差 分値に対して符号化処理を施し、 符号化した符号化差分値を上記被符号化画 素 P Xに対応する符号化画素値として出力するようになっている。 この符号 化処理では、 上記参照画素 P 0〜P 9の画素値から得られる予測画素値に基 づいて、 被符号化画素 P Xの画素値を符号化するための符号語を選択するよ うにしている。
次に動作について説明する。
第 2図は上記画像符号化装置 1 0 1の符号化処理をフローチャートにより 示しており、 まず、 このフローチャートに従って符号化処理の流れを簡単に 説明する。
画像信号が本画像符号化装置 1 0 1に入力されると、 上記画像信号を構成 する画素値が順次主メモリ 4に格納され、 該主メモリ 4からは、 被符号化画 素 P Xの符号化の際参照される、 被符号化画素 P Xの周辺に位置する複数の 参照画素 (周辺画素) P 0〜P 9の画素値が出力される (ステップ S 1 ) 。 次に、 上記主メモリ 4に取得された複数の参照画素のうちの最初の参照画 素が判定対象画素とされ (ステップ S 2 ) 、 この判定対象画素が符号化済画 素であるか否かの判定が符号化/未符号化判定器 1 0にて行われる (ステツ プ S 3 ) 。 この符号化/未符号化判定器 1 0では、 上記カウンタ出力 C o u tのインクリメントと、 上記主メモリ 4から 9つの参照画素 P 0〜P 9の画 素値が基準クロックに基づいて順次出力される動作とが同期しているため、 主メモリ 4から出力されている参照画素の、 被符号化画素 P Xに対する位置 を上記カウンタ出力 C 0 u tから検知することができる。
この判定の結果、 判定の対象となっている参照画素が符号化済画素であれ ば、 その画素値が参照画素値として上記予測値生成器 1 4にて取得され (ス テツプ S 5 ) 、 上記判定の対象となっている参照画素が符号化済画素でなけ れば、 この参照画素についてはその周辺の符号化済画素から参照画素値 (擬 似画素値) が生成され、 上記予測値生成器 1 4にて取得される (ステップ S 4 ) 。
次に、 被符号化画素を符号化する際に必要な参照画素値がすべて上記予測 値生成器 1 4にて取得できたか否かの判定がなされ (ステップ S 6 ) 、 必要 な参照画素値が全て取得されていない場合は、 上記主メモリ 4に取得された 参照画素のうちの次の参照画素が上記判定対象とされ (ステップ S 8 ) 、 上 記ステップ S 3〜S 6の処理が繰り返し行われる。 一方、 上記必要な参照画 素値が全て上記予測値生成器 1 4にて取得されている場合は、 上記予測値生 成器 1 4にて上記参照画素値に基づいて、 被符号化画素 P Xに対応する予測 画素値が生成される (ステップ S 7 ) 。
その後、 被符号化画素 P Xの画素値が上記符号化器 1 6 aに取得されると 同時に上記予測値生成器 1 4から被符号化画素 P Xの予測画素値が取得され (ステップ S 9 ) 、 上記符号化器 1 6 aにて、 上記被符号化画素 P Xの画素 値に対する符号化処理が上記予測画素値を用いて行われる (ステップ S 1 0 ) 。 次に、 画像信号の符号化処理における本画像符号化装置 1 0 1の動作とと もに、 上記各ステップ S 1〜S 1 0での本装置の各部の具体的な動作につい て詳しく説明する。
画像入力信号 I sが本画像符号化装置 1 0 1に入力されると、 該画像入力 信号 I sを構成する複数の画素はブロック化器 2にて、 1フレームを構成す る、 複数の画素からなるブロックに対応するようグループ化され、 各ブロッ ク内の画素に対応する画素値が符号化器 1 6 aに送られ、 該符号化器 1 6 a では、 被符号化画素 P Xの画素値を画素毎に参照画素値を参照して符号化す る符号化処理がプロック単位で行われる。
このとき、 上記主メモリ 4には、 走査線構造の画像入力信号 I sを構成す る画素値が順次格納され、 被符号化画素 P Xに対応する参照画素 P 0〜P 9 の画素値が、 一定の読出周期で出力される (ステップ S 1 ) 。 上記主メモリ 4の出力 Mは一時的に上記第 1 , 第 2補助メモリ 6 a, 6 bにて保持される。 上記第 1補助メモリ 6 aでは、 上記主メモリ 4の出力は 1読出周期の期間だ け、 上記第 2補助メモリ 6 bでは 2読出周期の期間だけ保持される。
また、 上記カウンタ 8では、 1フレームの最初の画素値を基準として、 画 像入力信号 I sに基づいて入力される画素値の数がカウントされ、 このカウ ント値 C 0 u tが符号化/未符号化判定器 1 0に出力される。 この符号化 Z 未符号化判定器 1 0では、 外部から 1フレームにおける縦横のブロック数情 報 B N sが入力されており、 このブロック数情報 B N s及ぴ上記カウンタ 8 の出力により指定される参照画素が、 上記符号化済画素か未符号化画素であ るかの判定がなされる判定対象画素となる。 例えば、 第 1 4図に示すように 被符号化画素 P Xが決まると、 これに対応する参照画素 P 0〜P 9が決まり、 このうちの主メモリ 4から最初に画素値が出力される画素 P 0がまず上記判 定対象画素となる (ステップ S 2 ) 。
このとき、 上記参照画素 P 0の画素値は、 上記第 1, 第 2補助メモリ 6 a, 6 bにでそれぞれ 1読出期間, 2読出期間の間保持される。 また、 符号化/未 符号化判定器 1 0は、 上記カウンタ出力 C 0 u t及びプロック数情報 B N s に基づいて、 被符号化ブロック内での被符号化画素の位置が計算され、 各参 照画素について符号化済画素であるか未符号化画素であるかが判定され (ス テツプ S 3 ) 、 この判定結果に応じて上記選択スィッチ 1 2が制御される。 上記参照画素 P 0は第 1 4図に示すように符号化済画素であるため、 符号 化/未符号化判定器 1 0による判定 (ステップ S 3 ) の結果、 選択スィッチ 1 2は該符号化/未符号化判定部 1 0により上記主メモリ 4の出力 Mを選択 するよう制御され、 これにより、 上記参照画素 P 0の画素値は参照画素値と して予測値生成器 1 4に格納される。
その後、 上記符号化ノ未符号化判定部 1 0では、 被符号化画素 P xに対応 する全ての参照画素の画素値が予測値生成器 1 4にて取得されたか否かが判 定される (ステップ S 6 ) 。 この場合は、 上記予測値生成器 1 4には参照画 素 P 0〜P 9の全ての参照画素値が取得されていないので、 上記符号化 Z未 符号化部 1 0では、 上記主メモリ 4から参照画素 P 0の画素値の次に出力さ れる参照画素 P 1の画素値が判定対象画素の画素値とされる (ステップ S 8 ) 。 そして、 参照画素 P 1は上記参照画素 P 0と同様に符号化済画素であるため、 この参照画素 P 1の画素値についても上記ステップ S 3 , S 5 , S 6 , S 8 の処理が行われる。
続いて、 上記符号化 Z未符号化部 1 0では、 上記主メモリ 4から参照画素 P 1の画素値の次に出力される参照画素 P 2の画素値が判定対象画素の画素 値とされる (ステップ S 8 ) 。 この参照画素 P 2は上記参照画素 P 0, P 1 とは異なり、 符号化済画素の隣に位置している未符号化画素であるため、 選 択スィツチ 1 2は該符号化 Z未符号化判定部 1 0により上記第 1補助メモリ 6 aの出力 M aを選択するよう制御され、 これにより、 上記参照画素 P 1の 画素値が、 参照画素 P 2の擬似画素値として予測値生成器 1 4に格納される (ステップ S 4 ) 。 その後は上記ステップ S 6 , ステップ S 8の処理が行わ れる。
さらに、 上記参照画素 P 3〜P 5については上記参照画素 P 0と同様に上 記ステップ S 3, S 5 , S 6 , S 8の処理が行われ、 さらに参照画素 P 6に ついては、 上記参照画素 P 2と同様にステップ S 3 , S 4 , S 6 , S 8の処 理が行われる。
続いて、 上記符号化 Z未符号化部 1 0では、 上記主メモリ 4から参照画素 P 6の画素値の次に出力される参照画素 P 7の画素値が判定対象画素の画素 値とされる (ステップ S 8 ) 。 この参照画素 P 7は上記参照画素 P 0〜P 6 とは異なり、 符号化済画素 P Xから 1画素分隔てて位置している未符号化画 素であるため、 選択スィツチ 1 2は該符号化/未符号化判定部 1 0により上 記第 2補助メモリ 6 bの出力 M bを選択するよう制御され、 これにより、 上 記参照画素 p 5の画素値が、 参照画素 P 7の擬似画素値として予測値生成器 1 4に格納される (ステップ S 4 ) 。 その後は上記ステップ S 6, ステップ S 8の処理が行われる。
さらに、 上記参照画素 P 8 , P 9については上記参照画素 P 0と同様に上 記ステップ S 3 , S 5 , S 6の処理が行われる。 この場合、 符号化/未符号 化部 1 0では、 全ての参照画素 P O〜P 9の全てについての参照画素値が上 記予測値生成器 1 4に取得されているという判定がなされ、 予測値生成器 1 4にて、 取得した参照画素値に基づいて、 被符号化画素 P Xに対応する予測 画素値 S pが算出される (ステップ S 7 ) 。
続いて、 上記符号化器 1 6 aには被符号化画素 P Xの画素値とともに上記 予測画素値 S pが取得され (ステップ S 9 ) 、 被符号化画素 P Xの画素値と その予測画素値 S pとの差分値に対して符号化処理が施されて、 符号化され た差分値が、 上記被符号化画素 P Xの符号化信号として出力される (ステツ プ S 1 0 ) 。 この符号化処理には、 上記予測画素値 S pに基づいて選択され た符号語が用いられる。
このようにして 1フレーム内の各ブロックの画素値が順次符号化されるこ ととなる。 なお、 1フレームの先頭画素については、 その周辺に符号化処理 が施された参照画素が存在しないので、 予測画素値は 0として符号化処理が 行われる。
このように本実施の形態 1では、未符号化画素の画素値を参照する場合は、 その周辺の符号化済画素の画素値を参照するようにしたので、 画素単位での 適応的な符号化処理とプロック単位での符号化処理とを、 未符号化画素と被 符号化画素との間での画素値の相関を損なうことなく、 しかも符号化信号の 復号化が困難となるのを回避しつつ組み合わせることができる。 これにより 伝送誤りの影響をブロック単位で収束可能とするとともに、 単純なブロック 単位の符号化処理に比べて符号化効率を向上することができ、 また複号化側 では、 被符号化画素の予測効率の劣化を招くことなく符号化された符号化信 号の復号化処理を正しく行うことが可能となる。 つまり、 符号化側にて被符号化画素の符号化処理をその周辺画素の画素値 を参照して行う場合は、 復号化側においても被復号化画素の復号化処理をそ の周辺画素の画素値を参照して行うこととなり、 しかも符号化処理の際に参 照される画素値と復号化処理の際に参照される画素値とを一致させる必要が ある。
そこで、 従来例の符号化方法では、 参照される画素が未符号化画素である 場合には、 その画素値として固定値を用いるようにしていたが、 この場合は 画素値の相関が損なわれるという問題があつた。
これに対し、 本発明では、 参照される画素が未符号化画素である場合には、 参照される符号化済画素から所定の規則で未符号化画素に対する擬似画素値 を生成するようにしている。 例えば、 第 1 5図の未符号化画素 P 2, P 6 , P 7についてはその画素値を符号化済画素 P 0, P 1 , P 3 , P 4 , P 5 , P 8 , P 9の画素値から一意的に決定するようにしている。 最も簡単な方法と しては、 未符号化画素と同じ水平走査線上に位置する、 距離が近い符号化済 画素の画素値を未符号化画素に対する擬似画素値としている。 この場合には、 未符号化画素 P 2の画素値は、 符号化済画素 P 1の画素値に置換され、 未符 号化画素 P 6, P 7の画素値は符号化済画素 P 5の画素値に置換される。
このようにすれば、 プロック境界での符号化済画素と未符号化画素の間で の画素値の相関が大きくなり、 画素間相関が大きいプロック内部と同じ予測 方式を用いて効率の高い符号化処理が実現できる。
このため、 画像信号をより低ビット数で画質を損なうことなく記録伝送す ることが可能となる。
なお、 上記実施の形態 1では、 主メモリ 4の出力 Mを保持する時間の異な る 2つの補助メモリ 6 a , 6 bを備えたが、 主メモリ 4の出力を一時的に保 持する補助メモリを 1つだけ備え、 主メモリ 4から出力される画素値を保持 する時間を、 未符号化画素の被符号化画素からの距離に応じて、 符号化 未 符号化判定器 1 0によって切り換えるようにしてもよい。
また、 上記実施の形態 1では、 画像信号の符号化処理の際、 被符号化画素 P Xの予測画素値のみを参照するものを示したが、 被符号化画素 P Xの予測 画素値だけでなく、 予測画素値の確からしさを示す予測確率に基づいて符号 化処理を行うようにしてもよレ、。
また、 この場合、 符号化処理の対象となる画像信号が 2値形状信号である 場合には、 予測画素値は 「0」 あるいは 「1」 のいずれかであるため、 予測 画素値をその一方に固定しておき、 予測確率のみを参照するようにしてもよ レ、。
以下、 本発明の実施の形態 2として、 上記予測画素値に代えて予測確率の みを参照して 2値形状信号の符号化処理を行う画像符号化装置を、 また実施 の形態 3として、 上記予測画素値とともに予測確率を参照して、 多値の画像 信号の符号化処理を行う画像符号化装置についてそれぞれ説明する。
実施の形態 2 .
第 3図は本発明の実施の形態 2による画像符号化装置 1 0 2の構成を示す ブロック図である。 なお、 図中、 第 1図と同一符号は上記実施の形態 1の画 像符号化装置 1 0 1におけるものと同一のものを示している。
この画像符号化装置 1 0 2は、 上記実施の形態 1の画像符号化装置 1 0 1 における予測値生成器 1 4に代えて予測確率生成器 2 2を備え、 画素値が 「0」 及び 「1」 のいずれかである 2値形状信号に対して符号化処理を行う よう構成したものである。
上記予測確率生成器 2 2は、 被符号化画素 P Xに対する参照画素 P 0〜P 9の画素値から、 被符号化画素 P Xの画素値とその予測画素値とがー致する 確率を予測確率として求め、 被符号化画素 P Xに対する予測確率信号 S kを 上記符号化器 1 6 bに出力する構成となっている。
ここで、 上記符号化器 1 6 bは、 本画像符号化装置 1 0 2が 2値形状信号 の符号化処理を対象するものであるため、 被符号化画素 P Xの画素値との引 き算を行う予測値は 「0」 及び 「1」 の一方に設定されている。 またこの符 号化器 1 6 bは、 上記予測確率が大きい場合は被符号化画素 P Xの画素値と その予測画素値とがー致する確率が高いので、 上記被符号化画素 P Xの画素 値とその予測画素値との差分値を、 差分値が 0となる場合に符号化効率が高 くなる符号化方法により符号化し、 一方、 上記予測確率が小さい場合は被符 号化画素 p xの画素値とその予測画素値とがー致する確率が小さいので、 上 記被符号化画素 P Xの画素値とその予測画素値との差分値を、 上記差分値が 0でない場合に符号化効率が高くなる符号化方法により符号化する構成とな つている。
次に作用効果について説明する。
なお、 上記実施の形態 1の画像符号化装置と共通する動作についてはその 説明を省略する。
このような構成の画像符号化装置 1 0 2による 2値形状信号の符号化処理 においても、 被符号化画素 P Xに対する参照画素 P 0〜P 9のうち未符号化 画素に相当するものについては、 その画素値が上記実施の形態 1の画像符号 化装置 1 0 1と同様に符号化済画素の画素値から生成され、 すべての参照画 素 P 0〜P 9に対応する画素値が予測確率生成器 2 2に格納される。
そして、 上記予測確率生成器 2 2では、 上記参照画素 P 0〜 P 9の画素値 に基づいて、 被符号化画素 P Xに対する予測確率を求める。 この予測確率情 報 S kが上記予測確率生成器 2 2から符号化器 1 6 bに出力されると、 該符 号化器 1 6 bでは、 被符号化画素 P Xの画素値と予め設定されている予測画 素値との差分値に対して、 上記予測確率情報 S kに応じた符号化処理を行う。 このように本実施の形態 2では、 2値形状信号の符号化処理を行う画像符 号化装置 1 0 2において、 ブロック境界に隣接する予測値生成が困難な被符 号化画素を、 符号化効率の劣化を抑えつつ符号化することができる。 このた め、 伝送誤りの影響をブロック単位で収束可能とするとともに、 単純なプロ ック単位の符号化処理に比べて符号化効率を向上することができ、 また復号 化側では、 被符号化画素の予測効率の劣化がなく符号化信号の複号化処理を 正しく行うことが可能となる。
なお、 予測確率に基づく符号化処理については、 例えば、 国際標準規格 J B I G (Joint Bi— level Image Coding Experts Group ) 【こ ¾ΰ載 eれてレヽる 力 この規格に記載されている符号化方法は、 符号化処理を画素単位 (非ブ ロック単位) で行うものであるため、 参照画素は常に符号化済画素であり、 本発明の解決課題であるブロック単位の符号化処理では参照画素が未符号化 画素となる場合の問題点はもとより、 この未符号化画素に対する画素値をど のように設定するかという対応策については何ら開示されていない。
実施の形態 3 .
第 4図は本発明の実施の形態 3による画像符号化装置 1 0 3の構成を示す ブロック図である。 なお、 図中、 第 1図と同一符号は上記実施の形態 1の画 像符号化装置 1 0 1におけるものと同一のものを示している。
この画像符号化装置 1 0 3は、 上述したように多値の画像信号に対して符 号化処理を行うものであり、 その予測確率生成部 1 3 0を上記実施の形態 1 の画像符号化装置 1 0 1における予測値生成器 1 4に加えて予測確率生成器 2 2を有する構成としたものである。
上記予測確率生成器 2 2は、 上記予測値生成器 1 4と同様、 選択スィッチ 1 2の出力 S o u tを受け、 被符号化画素 P Xに対する参照画素 P 0〜P 9 の画素値を順次格納し、 これらの画素値に基づいて、 上記予測値生成器 1 4 にて参照画素 P 0〜P 9に基づいて予測される予測画素値 S pの確からしさ を示す予測確率 S kを出力する構成となっている。
そして、 上記符号化器 1 6 cは、 上記予測値生成器 1 4からの予測画素値 S p及び予測確率生成器 2 2からの予測確率 S kに基づいて、 被符号化画素 P Xの画素値を符号化する構成となっている。
具体的には、 予測画素値の確からしさが大きい場合は、 被符号化画素の画 素値とその予測画素値の差が小さいので、 上記符号化器 1 6 cでは、 被符号 化画素の画素値とその予測画素値の差が小さい場合に符号化効率が高くなる 符号化方法により、 被符号化画素 P Xの画素値に対する符号化処理を行う。 また、 一方、 予測値の確からしさが小さい場合は、 被符号化画素の画素値と その予測画素値の差が大きいので、 符号化器 1 6 cでは、 画素値と予測画素 値との差分値がやや大きい場合に符号化効率が高い符号化方法により被符号 化画素 P Xの画素値に対する符号化処理を行う。
このような構成の画像符号化装置 1 0 3による多値の画像信号の符号化処 理においても、 被符号化画素 P Xに対する参照画素 P 0〜P 9のうち未符号 化画素に相当するものについては、 その画素値が上記実施の形態 1の画像符 号化装置 1 0 1と同様に符号化済画素の画素値から生成され、 すべての参照 画素 P 0〜P 9に対応する画素値が予測確率生成器 2 2及び予測値生成器 1 4に格納される。
そして、 予測値生成器 1 4では、 上記実施の形態 1と同様にして参照画素 P 0〜P 9から被符号化画素 P Xの予測画素値を求める。 また上記予測確率 生成器 2 2では、 上記参照画素 P 0〜P 9の画素値に基づいて、 被符号化画 素 P Xに対する予測確率 S kを求める。
上記予測画素値 S p及び予測確率 S kが符号化器 1 6 cに出力されると、 該符号化器 1 6 cでは、 被符号化画素 P Xの画素値と予測値生成器 1 4から の予測画素値との差分値に対して、 上記予測確率 S kに応じた符号化処理を 行う。
このように本実施の形態 3では、 多値の画像信号を符号化する画像符号化 装置 1 0 3において、 ブロック境界に隣接する予測値生成が困難な被符号化 画素を、 符号化効率の劣化を抑えつつ符号化することができる。 このため、 伝送誤りの影響をブロック単位で収束可能とするとともに、 単純なブロック 単位の符号化処理に比べて符号化効率を向上することができ、 また復号化側 では、 被符号化画素の予測効率の劣化がなく符号化信号の複号化処理を正し く行うことが可能となる。
実施の形態 4 .
第 5図は本発明の実施の形態 4による画像符号化装置 1 0 4の構成を示す ブロック図である。 なお、 図中、 第 1図と同一符号は上記実施の形態 1の画 像符号化装置 1 0 1におけるものと同一のものを示す。
この画像符号化装置 1 0 4は、 プロック化された画像信号 B sに対して非 可逆符号化処理を施す点で上記実施の形態 1の画像符号化装置 1 0 1とは異 なっている。
つまり、 上記画像符号化装置 1 0 4は、 上記実施の形態 1の可逆符号化処 理を行う符号化器 1 6 aに代えて、 上記プロック化器 2の出力 B sに対して D C T (離散コサイン変換) 処理を含む非可逆符号化処理を、 予測値生成器 1 4からの予測画素値に基づいて施す符号化器 1 6 dを有している。 さらに この画像符号化装置 1 0 4の予測値生成部 1 4 0は、 上記符号化器 1 6 dの 出力 C sを上記予測値生成器 1 4からの予測画素値 S pに基づいて復号化す る局所復号化器 2 4を備えており、 該局所複号化器 2 4の出力 L D sが主メ モリ 4に複号化済画素の画素値として格納され、 該出力 L D sが上記カウン タ 8に入力されるようになっている。 その他の構成は実施の形態 1の画像符 号化装置 1 0 1と同一である。
このような構成の画像符号化装置 1 0 4では、 画像入力信号 I sを被符号 化画素 P Xの予測画素値 S pに基づいて符号化処理する際、 該予測画素値 S pを生成する予測値生成部 1 4 0では、 局所複号化器 2 4にて符号化器 1 6 dの出力 C sを予測画素値 S pを参照して復号化し、 この復号化した画素値 を主メモリ 4に格納するようにしている。
このため、 非可逆符号化処理を行う画像符号化装置 1 0 4では、 予測画素 値の生成には復号化済画素値が使用されることとなり、 これによりこの画像 符号化装置により符号化された画像符号化信号を、 画像復号化装置で正しく 複号化することが可能となる。
なお、 上記実施の形態 4では、 画像信号の非可逆符号化処理の際、 被符号 化画素 P Xの予測画素値のみを参照するものを示したが、 被符号化画素 P X の予測画素値だけでなく、 予測画素値の確からしさを示す予測確率を参照す るようにしてもよレ、。
また、 この場合、 非可逆符号化処理の対象となる画像信号が 2値形状信号 である場合には、 予測画素値は 「0」 あるいは 「1」 のいずれかであるため、 予測画素値をその一方に固定しておき、 予測確率のみを参照するようにして もよい。
そこで、 本発明の実施の形態 4の変形例 1として、 上記予測画素値に代え て予測確率のみを参照して 2値形状信号の非可逆符号化処理を行う画像符号 化装置を、 また上記実施の形態 4の変形例 2として、 上記予測画素値ととも に予測確率を参照して、 多値の画像信号の非可逆符号化処理を行う画像符号 化装置についてそれぞれ説明する。
第 6図は本発明の実施の形態 4の変形例 1による画像符号化装置 1 0 4 a の構成を示すブロック図である。 なお、 図中、 第 5図と同一符号は実施の形 態 4の画像符号化装置 1 0 4におけるものと同一のものを示す。
この画像符号化装置 1 0 4 aは、 上記実施の形態 4の画像符号化装置 1 0 4における予測値生成器 1 4に代えて、予測確率発生器 2 2を備えたもので、 画素値が 「0」 及び 「1」 のいずれかである 2値形状信号の符号化処理を行 うものである。
上記予測確率生成器 2 2は、 被符号化画素 P Xに対する参照画素 P 0〜P 9の画素値から、 被符号化画素 P Xの画素値とその予測画素値とがー致する 確率を予測確率として求め、 被符号化画素 P Xに対する予測確率 S kを上記 符号化器 1 6 e及び局所復号化器 2 4に出力する構成となっている。
ここで、 上記符号化器 1 6 eは、 本画像符号化装置が 2値形状信号を処理 対象するものであるため、 被符号化画素 P xに対する予測画素値は 「0」 及 び 「1」 の一方に設定されている。 またこの符号化器 1 6 eは、 上記実施の 形態 2の符号化器 1 6 bと同様に、 上記予測確率が大きい場合は、 上記被符 号化画素 P Xの画素値とその予測画素値との差分値を、 差分値が 0となる場 合に符号化効率が高くなる符号化方法によって符号化し、 一方、 上記予測確 率が小さい場合は、 上記被符号化画素 P Xの画素値とその予測画素値との差 分値を、 上記差分値が 0でない場合に符号化効率が高くなる符号化方法によ つて符号化する構成となっている。
さらに上記局所複号化器 2 4は、 上記予測確率に基づいて上記符号化器 1 6 eと同様に複号化方法を切り替えて複号化処理を行うようになっている。 なお、 上記実施の形態 4の画像符号化装置と共通する動作についてはその 説明を省略する。
このような構成の画像符号化装置 1 0 4 aによる 2値形状信号の符号化処 理においても、 被符号化画素 P Xに対する参照画素 P 0〜P 9のうち未符号 化画素に相当するものについては、 その画素値が上記実施の形態 4の画像符 号化装置 1 0 4と同様に符号化済画素の画素値から生成され、 すべての参照 画素 P 0〜P 9に対応する画素値が予測確率生成器 2 2に格納される。
そして、 上記予測確率生成器 2 2では、 上記参照画素 P 0〜P 9の画素値 に基づいて、 被符号化画素 P xに対する予測確率 S kを求める。 この予測確 率 S kが上記予測確率生成器 2 2から符号化器 1 6 e及び局所復号化器 2 4 に出力されると、 該符号化器 1 6 eでは、 被符号化画素 P Xの画素値と予め 設定されている予測画素値との差分値に対して、 上記予測確率 S kに応じた 非可逆符号化処理を行う。 この際、 上記局所復号化器 2 4では、 上記予測確 率 S kに応じた、 上記符号化器 1 6 eの出力 C sに対する復号化処理を行つ て、 被符号化画素 P Xの画素値を再生する。 この再生された被符号化画素 P Xの画素値は上記主メモリ 4に格納される。
このように本実施の形態 4の変形例 1では、 2値形状信号に対する非可逆 符号化処理を行う画像符号化装置 1 0 4 aにおいて、 ブロック境界に隣接す る予測値生成が困難な被符号化画素を、 符号化効率の劣化を抑えつつ符号化 することができる。 このため、 伝送誤りの影響をブロック単位で収束可能と するとともに、 単純なプロック単位の符号化処理に比べて符号化効率を向上 することができ、 また複号化側では、 被符号化画素の予測効率の劣化がなく 符号化信号の復号化処理を正しく行うことが可能となる。
第 7図は本発明の実施の形態 4の変形例 2による画像符号化装置 1 0 4 b の構成を示すブロック図である。 なお、 図中、 第 4図と同一符号は上記実施 の形態 4の画像符号化装置 1 0 4におけるものと同一のものを示す。
この画像符号化装置 1 0 4 bは、 上述したように多値の画像信号に対して 非可逆符号化処理を行うものであり、 その予測確率生成部 1 4 0 bを上記実 施の形態 4の画像符号化装置 1 0 4における予測値生成器 1 4に加えて予測 確率生成器 2 2を有する構成としたものである。
上記予測確率生成器 2 2は、 上記予測値生成器 1 4と同様、 選択スィッチ 1 2の出力 S 0 u tを受け、 被符号化画素 P Xに対する参照画素 P 0〜P 9 の画素値を順次格納し、 これらの画素値に基づいて、 上記予測値生成器 1 4 にて参照画素 P 0〜P 9に基づいて予測される予測画素値 S pの確からしさ を示す予測確率 S kを出力する構成となっている。
そして、 上記符号化器 1 6 f は、 上記予測値生成器 1 4からの予測画素値 S p及び予測確率生成器 2 2からの予測確率 S kに基づいて、 被符号化画素 P xの画素値に対して非可逆符号化処理を施す構成となっており、 その具体 的な構成は、 上記実施の形態 3の符号化器 1 6 cと全く同一となっている。 さらに上記局所復号化器 2 は、 上記予測確率に基づいて上記符号化器 1 6 eと同様に符号化方法を切り替えて復号化処理を行うようになっている。 このような構成の画像符号化装置 1 0 4 bによる多値の画像信号の符号化 処理においても、 被符号化画素 P Xに対する参照画素 P 0〜P 9のうち未符 号化画素に相当するものについては、 その画素値が上記実施の形態 4の画像 符号化装置 1 0 4と同様に符号化済画素の画素値から生成され、 すべての参 照画素 P 0〜P 9に対応する画素値が予測確率生成器 2 2及び予測値生成器 1 4に格納される。
そして、 予測値生成器 1 4では、 上記実施の形態 4と同様にして参照画素 P 0〜P 9から被符号化画素 P Xの予測画素値を求める。 また上記予測確率 生成器 2 2では、 上記参照画素 P 0〜P 9の画素値に基づいて、 被符号化画 素 P Xに対する予測確率 S kを求める。
上記予測画素値 S p及び予測確率 S kが符号化器 1 6 f 及び局所復号化器 2 4に出力されると、 該符号化器 1 6 f では、 被符号化画素 P Xの画素値と 予め設定されている予測画素値との差分値に対して、 上記予測確率 S kに応 じた符号化処理を行う。 この際、 上記局所復号化器 2 4では、 上記予測確率 S kに応じた、 上記符号化器 1 6 f の出力 C sに対する複号化処理を行って、 被符号化画素 P Xの画素値を再生する。 この再生された被符号化画素 P Xの 画素値は上記主メモリ 4に格納される。
これにより予測値生成が困難な画像信号に対しては符号化効率を劣化させ ず、 予測値生成が容易な画像信号に対しては大幅な符号化効率向上を達成す ることができる。
このように本実施の形態 4の変形例 2では、 多値の画像信号を非可逆符号 化する画像符号化装置 1 0 4 bにおいて、 ブロック境界に隣接する予測値生 成が困難な被符号化画素を、 符号化効率の劣化を抑えつつ符号化することが できる。 このため、 伝送誤りの影響をブロック単位で収束可能とするととも に、 単純なプロック単位の符号化処理に比べて符号化効率を向上することが でき、 また復号化側では、 被符号化画素の予測効率の劣化がなく符号化信号 の複号化処理を正しく行うことが可能となる。
実施の形態 5 .
第 8図は本発明の実施の形態 5による画像複号化装置 1 0 5の構成を示す ブロック図である。
この実施の形態 5の画像複号化装置 1 0 5は、 実施の形態 1の画像符号化 装置 1 0 1により可逆符号化処理を施した画像符号化信号を可逆的に複号化 するものである。
この画像複号化装置 1 0 5は、 入力された画像符号化信号 C sを、 1画像 表示領域を構成する複数のブロック毎に復号化し、 この際、 復号化処理の対 象となる被復号化画素 P χ ' の画素値, つまり被符号化画素 P Xの画素値を 符号化した符号化信号を、 被復号化画素の予測画素値に基づいて複号化する 復号化器 2 6 aと、 該復号化器 2 6 aの出力である各プロックに対応した画 像複号化信号 D sを統合して所定の走査線構造の画像再生信号 R sを生成す る逆ブロック化器 3 0と、 上記予測画素値を上記被復号化画素 P x ' の周辺 に位置する参照画素 P 0 ' ~ Ρ 9 ' の画素値に基づいて生成する予測値生成 部 1 5 0とを有している。
ここで、 上記予測値生成部 1 5 0は、 実施の形態 1の画像符号化装置 1 0 1における予測値生成部 1 1 0とほぼ同様な構成となっている。
すなわち、 該予測値生成部 1 5 0は、 例えば 1フレームに相当する数だけ 記憶可能な大容量の主メモリ 4と、 該主メモリ 4から出力される画素値 Μを 一時的に保持する保持時間の異なる第 1, 第 2補助メモリ 6 a, 6 bとを有 している。 ここで、 上記主メモリ 4は、 1つの画素値の符号化処理に要する 時間の間に、 上記復号化器 2 6 aにて処理の対象となっている被復号化画素 Ρ χ ' の参照画素 Ρ Ο ' 〜Ρ 9 ' (第 1 7図参照) に対応する画素値を、 上 記記憶した画素値から順次出力する構成となっている。 また、 上記第 1補助 メモリ 6 aは、 主メモリ 4から順次出力される画素値 Mを 1画素分だけ遅延 する構成となっており、 上記第 2補助メモリ 6 bは主メモリ 4から順次出力 される画素値 Mを 2画素分だけ遅延する構成となっている。 また、 上記予測値生成部 1 5 0は、 各フレームの画像復号化信号 D sを受 け、 画素値の数をカウントするカウンタ 8と、 該カウンタ 8の出力 C o u t 及び外部から供給される各フレームにおける縦横のプロック数の情報 B N s 基づいて、 主メモリ 4から出力されている画素値が、 すでに符号化器 2 6 a にて復号化処理が施された復号化済画素の画素値であるか、 まだ復号化処理 が施されていない未復号化画素の画素値であるかを判定する復号化/未復号 化判定器 2 0とを有している。 ここで該判定器 2 0は、 未復号化画素につい ては、 これと同一水平走査線上に位置する最も近い複号化済画素までの距離 を画素数でもって計測する処理も行っている。 また、 上記カウンタ 8は、 1 フレームを構成するすべての画素の画素値が入力された時点でリセットされ るようになっている。
さらに、 上記予測値生成部 1 5 0は、 上記復号化/未復号化判定器 2 0の 出力 S c o n tに基づいて上記主メモリ 4 , 第 1, 第 2補助メモリ 6 a, 6 bのいずれかの出力 M, M a , M bを選択して出力する選択スィッチ 1 2と、 該選択スィッチ 1 2の出力 S o u tを、 被復号化画素 P x ' に対応する予測 画素値の生成に必要な参照画素 P 0 ' 〜Ρ 9 ' の画素値として取得し、 該被 複号化画素 Ρ χ ' に対応する予測画素値 S ρを生成する予測値生成器 1 4と を有している。
そして、 この画像復号化装置 1 0 5では、 上記復号化器 2 6 aは、 外部か ら画像符号化信号として入力される被符号化画素 P Xに対する符号化差分値 を復号化して複号化差分値を生成し、 該複号化差分値に上記予測値生成部 1 5 0からの予測画素値 S pを加算して被復号化画素に対する復号化画素値を 生成して逆プロック化器 3 0に出力するようになっている。
次に動作について説明する。
第 9図は上記画像復号化装置 1 0 5の複号化処理をフローチャートにより 示しており、 まず、 このフローチャートに従って複号化処理の流れを簡単に 説明する。
画像符号化信号 C sが本画像符号化装置 1 0 5に入力され、 復号化器 2 6 aにて該画像符号化信号 C sの復号化処理が予測値生成部 1 5 0からの予測 信号 S pに基づいて行われる。
このとき上記複号化器 26 aの出力として、 1フレームに対応する複数の 複号化画素値が順次主メモリ 4に格納され、 該主メモリ 4からは、 被複号化 画素 Ρ χ の複号化の際に参照される、 被復号化画素 Ρχ' の周辺に位置す る複数の参照画素 (周辺画素) Ρ Ο' 〜Ρ 9' の画素値が出力される (ステ ップ S 11 ) 。
次に、 上記主メモリ 4に取得された複数の参照画素のうちの最初の参照画 素が判定対象画素とされ (ステップ S 12) 、 この判定対象画素が復号化済 画素であるか否かの判定が復号化/未複号化判定器 30にて行われる (ステ ップ S 1 3) 。 この符号化/未符号化判定器 30では、 上記カウンタ出力 C o u tのインクリメントと、 上記主メモリ 4から 9つの参照画素 P 0' 〜P 9' の画素値が基準ク口ックに基づいて順次出力される動作とが同期してい るため、 主メモリ 4から出力されている参照画素の、 被複号化画素 Ρ χ' に 対する位置を上記カウンタ出力 C o u tから検知することができる。
この判定の結果、 判定の対象となっている参照画素が復号化済画素であれ ば、 その画素値が参照画素値として上記予測値生成器 14にて取得され (ス テツプ S 15) 、 上記判定の対象となっている参照画素が復号化済画素でな ければ、 この参照画素についてはその周辺の復号化済画素から参照画素値 (擬 似画素値) が生成され、 上記予測値生成器 14にて取得される (ステップ S 14) 。
次に、 被復号化画素を複号化する際に必要な参照画素値がすべて上記予測 値生成器 14にて取得できたか否かの判定がなされ (ステップ S 16) 、 必 要な参照画素値が全て取得されていない場合は、 上記主メモリ 4に取得され た参照画素のうちの次の参照画素が上記判定対象とされ (ステップ S 18) 、 上記ステップ S 1 3〜S 16の処理が繰り返し行われる。 一方、 上記必要な 参照画素値が全て上記予測値生成器 14にて取得されている場合は、 上記予 測値生成器 14にて上記参照画素値に基づいて、 被復号化画素 P x' に対応 する予測画素値が生成される (ステップ S 17) 。
その後、 被復号化画素 P χ' の画素値が上記復号化器 26 aに取得される と同時に上記予測値生成器 14から被復号化画素 P x' の予測画素値が取得 され (ステップ S 19) 、 上記複号化器 26 aにて、 上記被復号化画素 P χ' の画素値に対する復号化処理が上記予測画素値を用いて行われる (ステップ
520) 。
次に、 画像信号の複号化処理における本画像複号化装置 105の動作とと もに、 上記各ステップ S 1 1〜S 20での本装置の各部の具体的な動作につ いて詳しく説明する。
画像符号化信号 C sが本画像復号化装置 105に入力されると、 該画像復 号化信号 C sは、 復号化器 26 aに送られ、 該復号化器 26 aでは、 被復号 化画素 Ρ χ' の画素値を画素毎に参照画素値を参照して複号化する復号化処 理がブロック単位で行われる。
このとき、 上記主メモリ 4には、 上記複号化器 26 aの出力である 1フレ ームを構成する画素値が順次格納され、 被復号化画素 P x' に対応する参照 画素 P 0' 〜Ρ 9' の画素値が、 一定の読出周期で出力される (ステップ S 11) 。 上記主メモリ 4の出力 Μは一時的に上記第 1, 第 2補助メモリ 6 a,
6 bにて保持される。 上記第 1補助メモリ 6 aでは、 上記主メモリ 4の出力 は 1読出周期の期間だけ、 上記第 2補助メモリ 6 bでは 2読出周期の期間だ け保持される。
また、 上記カウンタ 8では、 上記 1フレームの最初の画素値を基準として、 複号化器 26 aの出力 D sに基づいて、 入力される画素値の数がカウントさ れ、 このカウント値 C o u tが復号化/未復号化判定器 20に出力される。 この復号化/未復号化判定器 20では、 外部から 1フレームにおける縦横の ブロック数情報 BN sが入力されており、 このブロック数情報 BN s及び上 記カウンタ 8の出力により指定される参照画素が、 上記復号化済画素か未復 号化画素であるかの判定がなされる判定対象画素となる。 例えば、 第 1 7図 に示すように被複号化画素 P χ' が決まると、 これに対応する参照画素 P 0' 〜Ρ 9' が決まり、 このうちの主メモリ 4から最初に画素値が出力される画 素 P0' がまず上記判定対象画素となる (ステップ S 12) 。
このとき、 上記参照画素 P 0' の画素値は、 上記第 1, 第 2補助メモリ 6 a, 6 bにでそれぞれ 1読出期間, 2読出期間の間保持される。 また、 復号 化 Z未複号化判定器 20は、 上記カウンタ出力 C o u t及びブロック数情報 BN sに基づいて、 被複号化プロック内での被複号化画素の位置が計算され、 各参照画素について複号化済画素であるか未復号化画素であるかが判定され、 この判定結果に応じて上記選択スィッチ 12が制御される。
この画素 P 0' は第 1 7図に示すように複号化済画素であるため、 複号化 Z未複号化判定器 20による判定 (ステップ S 13) の結果、 上記選択スィ ツチ 12は該復号化/未複号化判定器 20により上記主メモリ 4の出力 Mを 選択するよう制御され、 これにより、 上記画素 P 0' の画素値は参照画素値 として予測値生成器 14に格納される。
その後、 上記復号化/未復号化判定器 20では、 被復号化画素 P x' に対 応する全ての参照画素の画素値が予測値生成器 14にて取得されたか否かを 判定する (ステップ S 16) 。 この場合は、 上記予測値生成器 14には参照 画素 P 0' 〜Ρ 9' の全ての参照画素値が取得されていないので、 上記復号 化 Ζ未復号化判定器 20は、 上記主メモリ 4から参照画素 Ρ 0' の画素値の 次に出力される参照画素 Ρ 1' の画素値を判定対象画素の画素値とする (ス テツプ S 18) 。 そして、 参照画素 P 1' は上記参照画素 P 0' と同様に復 号化済画素であるため、 この参照画素 P 1' の画素値についても上記ステツ プ S 13, S 15, S 16, S 18の処理が行われる。
続いて、 上記複号化/未複号化判定器 20では、 上記主メモリ 4から参照 画素 P 1' の画素値の次に出力される参照画素 Ρ 2' の画素値が判定対象画 素の画素値とされる (ステップ S 18) 。 この参照画素 Ρ 2' は上記参照画 素 P 0' , P 1' とは異なり、 復号化済画素の隣に位置している未復号化画 素であるため、 上記選択スィッチ 12は該複号化 Ζ未復号化判定器 20によ り上記第 1補助メモリ 6 aの出力 Maを選択するよう制御され、 これにより、 上記参照画素 P 1' の画素値が、 参照画素 P 2' の擬似画素値として予測値 生成器 14に格納される (ステップ S 14) 。 その後は上記ステップ S 16, ステップ S 18の処理が行われる。
さらに、 上記参照画素 P 3' 〜Ρ 5' については上記参照画素 Ρ 0' と同 様に上記ステップ S I 3, S I 5, S I 6, S I 8の処理が行われ、 さらに 参照画素 p 6' については、 上記参照画素 P 1 ' と同様にステップ S 13' , S 1 ' , S 16' , S 18' の処理が行われる。
続いて、 上記複号化 Z未復号化判定器 20では、 上記主メモリ 4から参照 画素 P 6' の画素値の次に出力される参照画素 P 7' の画素値が判定対象画 素の画素値とされる (ステップ S 18) 。 この参照画素 は上記参照画 素 P 0' 〜Ρ 6' とは異なり、 復号化済画素から 1画素分隔てて位置してい る未復号化画素であるため、 上記選択スィッチ 12は該復号化 Ζ未復号化判 定器 20により上記第 2補助メモリ 6 bの出力 Mbを選択するよう制御され、 これにより、 上記参照画素 P 5' の画素値が、 参照画素 P 7' の擬似画素値 として予測値生成器 14に格納される (ステップ S 14) 。 その後は上記ス テツプ S 16, ステップ S 18の処理が行われる。
さらに、 上記参照画素 P 8' , P 9' については上記参照画素 P と同 様に上記ステップ S 13, S 15, S 16の処理が行われる。 この場合、 復 号化/未複号化判定器 20は全ての参照画素 P 0' 〜Ρ 9' の全てについて の参照画素値が上記予測値生成器 14に取得されているという判定がなされ、 予測値生成器 14にて、 取得した参照画素値に基づいて、 被復号化画素 Ρχ' に対応する予測画素値が算出される (ステップ S 17) 。
続いて、 上記復号化器 26 aには被復号化画素 Ρ χ' の画素値とともに上 記予測画素値が取得され (ステップ S 19) 、 被複号化画素 P x' の画素値 とその予測画素値との加算値が、 上記被複号化画素 P Xの画像複号化信号 D sとして出力される (ステップ S 20) 。
このようにして 1フレーム内の各ブロックの画素値が復号化処理される。 なお、 1フレームの先頭画素については、 復号化処理を施した参照画素が存 在しないので、 予測画素値を 0として復号ィヒ処理が行われる。
そして上記画像複号化信号 D sは逆プロック器 30にて 1フレーム画面に 対応するよう統合されて、 走査線構造の画像再生信号 R sが出力される。
このように本実施の形態 5では、未複号化画素の画素値を参照する場合は、 その周辺の符号化済画素の画素値を参照するようにしたので、 画素単位での 適応的な復号化処理とプロック単位での複号化処理とを、 未復号化画素と被 復号化画素との間での画素値の相関を損なうことなく行うことができる。 こ れにより伝送誤りの影響をプロック単位で収束可能とするとともに、 単純な プロック単位の符号化処理に比べて符号化効率を向上した符号化方法により 処理された画像符号化信号 C sを正しく復号化することができる。
なお、 上記実施の形態 5では、 主メモリ 4の出力 Mを保持する時間の異な る 2つの補助メモリ 6 a, 6 bを備えたが、 主メモリ 4の出力を一時的に保 持する補助メモリを 1つだけ備え、 主メモリ 4から出力される画素値を保持 する時間を、 未複号化画素の被復号化画素からの距離に応じて、 複号化 Z未 復号化判定器 2 0によって切り換えるようにしてもよい。
また、 上記実施の形態 5では、 画像信号の符号化処理の際、 被複号化画素 P χ ' の予測画素値のみを参照するものを示したが、 被復号化画素 Ρ χ ' の 予測画素値だけでなく、 予測画素値の確からしさを示す予測確率に基づいて 復号化処理を行うようにしてもよい。
また、 この場合、 複号化処理の対象となる画像信号が 2値形状信号である 場合には、 予測画素値は 「0」 あるいは 「1」 のいずれかであるため、 予測 画素値をその一方に固定しておき、 予測確率のみを参照するようにしてもよ レ、。
以下、 本発明の実施の形態 6として、 上記予測画素値に代えて予測確率の みを参照して 2値形状信号の複号化処理を行う画像復号化装置を、 また実施 の形態 7として、 上記予測画素値とともに予測確率を参照して、 多値の画像 信号の復号化処理を行う画像復号化装置についてそれぞれ説明する。
実施の形態 6 .
第 1 0図は本発明の実施の形態 6による画像複号化装置の構成を示すプロ ック図である。 なお、 図中、 第 8図と同一符号は上記実施の形態 5の画像復 号化装置 1 0 5におけるものと同一のものを示す。
本実施の形態 6の画像複号化装置 1 0 6は、 上記実施の形態 5の画像復号 化装置 1 0 5における予測値生成器 1 4に代えて予測確率生成器 2 2を備え、 画素値が 「0」 及び 「1」 のいずれかである 2値形状信号に対する複号化処 理を行うようにしたものである。
上記予測確率生成器 22は、 被復号化画素 Ρ χ' に対する参照画素 Ρ Ο' 〜Ρ 9' の画素値から、 被複号化画素 Ρ χ' の画素値とその予測画素値とが 一致する確率を予測確率 S kとして求め、 被複号化画素 P x' に対する予測 確率 S kを上記複号化器 26 bに出力する構成となっている。
ここで、 上記復号化器 26 bは、 上記実施の形態 2の画像符号化装置 10 2により符号化された画像符号化信号 C sの復号化処理を行う構成となって いる。
次に作用効果について説明する。
なお、 上記実施の形態 1の画像符号化装置と共通する動作についてはその 説明を省略する。
このような構成の画像復号化装置 105による 2値形状信号の復号化処理 においても、 被復号化画素 P x' に対する参照画素 P 0' 〜Ρ 9' のうち未 複号化画素に相当するものについては、 その画素値が上記実施の形態 5の画 像復号化装置 105と同様に復号化済画素の画素値から生成され、 すべての 参照画素 P C) 〜Ρ 9' に対応する画素値が予測確率生成器 22に格納され る。
そして、 上記予測確率生成器 22では、 上記参照画素 Ρ Ο' 〜Ρ 9' の画 素値に基づいて、 被複号化画素 Ρ χ' に対する予測確率 S kを求める。 この 予測確率 S kが上記予測確率生成器 22から複号化器 26 bに出力されると、 該複号化器 26 bでは、 被復号化画素 P x' の画素値と予め設定されている 予測画素値との差分値に対して、 上記予測確率 S kに応じた複号化処理を行 5。
このように本実施の形態 6では、 2値の画像信号を符号化してなる画像符 号化信号を復号化する画像復号化装置 106において、 ブロック境界に隣接 する予測値生成が困難な被符号化画素を、 符号化効率の劣化を抑えつつ符号 化することができる符号化処理に対応する復号化処理を実現できる。
実施の形態 7.
第 1 1図は本発明の実施の形態 7による画像復号化装置の構成を示すプロ ック図である。 なお、 図中、 第 8図と同一符号は上記実施の形態 5の画像符 号化装置 105におけるものと同一のものを示す。
本実施の形態 7の画像復号化装置 107は、 上述したように多値の画像信 号に対して符号化処理を行うもので、 上記実施の形態 3の画像符号化装置 1 03により符号化された画像符号化信号 C sの復号化処理を行う構成となつ ている。
具体的には、 上記画像複号化装置 107の予測確率生成部 170は、 上記 実施の形態 5の画像符号化装置 105における予測値生成器 14に加えて予 測確率生成器 22を有する構成となっている。
上記予測確率生成器 22は、 上記予測値生成器 14と同様、 選択スィツチ 12の出力 S o u tを受け、 被複号化画素 P に対する参照画素 P 0' 〜 P 9' の画素値を順次格納し、 これらの画素値に基づいて、 上記予測値生成 器 14にて参照画素 P0' 〜Ρ 9' に基づいて予測される予測画素値 S ρの 確からしさを示す予測確率 S kを出力する構成となっている。
そして、 上記復号化器 26 cは上記実施の形態 3の符号化器 16 cによる 符号化処理に対応する復号化処理を行うものであり、 該複号化器 26 cは、 上記予測値生成器 14からの予測画素信号 S 及び予測確率生成器 22から の予測確率 S kに基づいて、 被復号化画素 Ρ χ' の画素値を複号化する構成 となっている。
次に作用効果について説明する。
なお、 上記実施の形態 5の画像符号化装置と共通する動作についてはその 説明を省略する。
このような構成の画像復号化装置 107による多値の画像信号の復号化処 理においても、 被複号化画素 Ρ χ' に対する参照画素 Ρ Ο' 〜Ρ 9' のうち 未複号化画素に相当するものについては、 その画素値が上記実施の形態 5の 画像復号化装置 105と同様に復号化済画素の画素値から生成され、 すべて の参照画素 Ρ Ο' 〜Ρ 9' に対応する画素値が予測確率生成器 22及び予測 値生成器 14に格納される。
そして、 予測値生成器 14では、 上記実施の形態 5と同様にして参照画素 P 0' 〜P 9 から被復号化画素 Ρ χ' の予測画素値を求める。 また上記予 測確率生成器 22では、 上記参照画素 P CT 〜Ρ 9' の画素値に基づいて、 被復号化画素 Ρχ' に対する予測確率を求める。
上記予測画素値 S ρ及び予測確率 S kが復号化器 26 c及び局所復号化器 24に出力されると、 該複号化器 26 cでは、 被複号化画素 P x' の画素値 と予測値生成器 14からの予測画素値との加算値に対して、 上記予測確率 S kに応じた複号化処理を行う。 この再生された被復号化画素 P X ' の画素値 は上記主メモリ 4に格納される。
このように本実施の形態 7では、 多値の画像信号を符号化してなる画像符 号化信号を復号化する画像復号化装置 107において、 ブロック境界に隣接 する予測値生成が困難な被符号化画素を、 符号化効率の劣化を抑えつつ符号 化することができる符号化処理に対応する複号化処理を実現できる。
なお、 上記実施の形態 5, 6, 7では、 画像複号化装置として、 実施の形 態 1, 2, 3の画像復号化装置 101, 102, 103により可逆符号化し た画像符号化信号を複号化するものを示したが、 上記復号化器 26 aを、 非 可逆符号化に対応する複号化処理を行う構成とすることにより、 上記実施の 形態 5, 6, 7における画像複号化装置 105, 106, 107を、 実施の 形態 4, その変形例 1, 2の画像符号化装置 104, 10 a, 104 bに 対応するものとできる。
また、 上記実施の形態で示した符号化装置あるいは復号化装置の構成を実 現するための符号化あるいは複号化プログラムを、 フロッピーディスク等の データ記憶媒体に記録するようにすることにより、 上記実施の形態で示した 処理を、 独立したコンピュータシステムにおいて簡単に実施することが可能 となる。
第 12図(a) は、 上記実施の形態の符号化あるいは複号化処理を、 上記符 号化あるいは復号化プログラムを格納したフロッピーディスクを用いて、 コ ンピュータシステムにより実施する場合を説明するための図である。
第 12図(b) は、 フロッピーディスクの正面からみた外観、 断面構造、 及 びフロッピーディスクを示し、 第 12図(a) は、 記録媒体本体であるフロッ ピーディスクの物理フォーマツ トの例を示している。 フロッピーディスク F Dはケース F内に内蔵され、 該ディスクの表面には、 同心円状に外周からは 内周に向かって複数のトラック T rが形成され、 各トラックは角度方向に 1 6のセクタ S eに分割されている。 従って、 上記プログラムを格納したフロ ッピーデイスクでは、 上記フロッピーデイスク F D上に割り当てられた領域 に、 上記プログラムとしてのデータが記録されている。
また、 第 1 2図(c) は、 フロッピーディスク F Dに上記プログラムの記録 再生を行うための構成を示す。 上記プログラムをフロッピーディスク F Dに 記録する場合は、 コンピュータシステム C sから上記プログラムとしてのデ ータをフロッピーディスク ドライブを介して書き込む。 また、 フロッピーデ イスク内のプログラムにより上記符号化あるいは復号化装置をコンピュータ システム中に構築する場合は、 フロッピーディスク ドライブによりプロダラ ムをフロッピーディスクから読み出し、 コンピュータシステムに転送する。 上記説明では、 データ記録媒体としてフロッピーディスクを用いて説明を 行ったが、 光ディスクを用いても同様に行うことができる。 また、 記録媒体 はこれに限らず、 I Cカード、 R OMカセット等、 プログラムを記録できる ものであれば同様に実施することができる。
また、 上記実施の形態 2, 3及び実施の形態 4の変形例 1, 2では、 予測 確率に応じて符号化方法を切換える例を、 実施の形態 5, 6では、 予測確率 に応じて複号化方法を切換える例を示したが、 上記予測確率に応じて符号語 (符号化テーブル) を変更してもよい。 特に算術符号で符号化する場合は予 測確率により算術符号に相当する確率テーブルを更新するようにすることに より、 簡単な構成により、 上記実施の形態 2, 3及び実施の形態 4の変形例 1, 2の画像符号化装置、 あるいは実施の形態 5, 6の画像複号化装置を実 現することができ、 この場合、 その実用上の効果は非常に大きい。
また、 本発明は、 被符号化画素の予測画素値をその周辺に位置する複数の 周辺画素の画素値を参照して予測する際、 該周辺画素のうちの未符号化画素 については、 上記周辺画素のうちの符号化済画素の画素値を用いて擬似画素 値を生成するようにしたものであるが、 被符号化画素のプロック内での位置 に応じて、 その予測画素値を生成する際に参照される 1群の周辺画素として、 被符号化画素周辺での配置が異なるものを用いるようにしてもよい。
例えば、 第 14図を用いて具体的に説明すると、 被符号化画素 P xがプロ ックの境界に位置する場合には、 その周辺画素 P 0, P 1, P 3, P 4, P 5, P 8, P 9のみを参照画素とし、 被符号化画素 P Xがブロックの境界以 外に位置する場合には、 その周辺画素 PO, P 1, P 2, P 3, P 4, P 5, P 6, P 7, P 8, P 9の全てを参照画素とする。 そして、 被符号化画素 P X を符号化する際には、 その周辺画素 P O, P 1, P 3, P 4, P 5, P 8, P 9のみで構成される符号語と、 周辺画素 P 0, P 1, P 2, P 3, P 4, P 5, P 6, P 7, P 8, P 9の全てで構成される符号語とを切り替えて使用 する。
言い換えると、 符号化器を、 被符号化画素 P Xのブロック内での位置に応 じた符号語を複数有する構成とし、 上記被符号化画素 P Xの位置に応じて符 号語を切換える構成とする。
このような構成によっても、 上述した本発明の実施の形態の画像符号化装 置と同様な効果が得られることは、 本発明の実施の形態から容易に類推でき る。
なお、 復号化器についても、 第 17図に示す被復号化画素 Px' のブロッ ク内での位置に応じた符号語を複数有する構成とし、上記被復号化画素 P の位置に応じて符号語を切換える構成とすることにより、 上記実施の形態の 画像複号化装置と同様な効果が得られることは言うまでもない。 産業上の利用可能性
以上のように本発明に係る画像符号化装置及び画像符号化方法, 画像復号 化装置及び画像複号化方法, 並びにデータ記録媒体は、 画像信号の圧縮処理 における符号化効率の向上を図ることができ、 画像信号の伝送や記憶を行う システムにおける画像符号化処理や画像複号化処理を実現するものとして極 めて有用であり、 特に、 MP EG 4等の規格に準拠した動画像の圧縮, 伸長 処理に適している。

Claims

請 求 の 範 囲
1 . 画像信号を構成する画素値を、 被符号化画素の周辺に位置する複数の 周辺画素の画素値に基づいて順次符号化する画像符号化装置であって、
1画像表示領域に対応する複数の画素値からなる画像信号を、 所定数の画 素からなるプロックに対応するようプロック化し、 各プロックの画像信号を 構成する所定数の画素値をブロック単位で出力するブロック化手段と、 上記周辺画素のうちの未符号化画素の画素値を、 上記周辺画素のうちの符 号化済画素の画素値から所定の規則に基づいて得られる擬似画素値と置換す る画素値置換手段と、
上記各プロックに対応する複数の画素値からなる画像信号を受け、 該各画 素値を上記符号化済画素の画素値及び未符号化画素の擬似画素値に基づいて 順次符号化する符号化処理を各ブロック毎に行って画像符号化信号を出力す る符号化手段とを備えたことを特徴とする画像符号化装置。
2 . 請求の範囲第 1項記載の画像符号化装置において、
上記画素値置換手段は、 上記未符号化画素からの空間距離が最短である符 号化済画素の画素値を該未符号化画素の擬似画素値とするものであることを 特徴とする画像符号化装置。
3 . 請求の範囲第 1項記載の画像符号化装置において、
上記画素値置換手段は、 上記未符号化画素からの空間距離が最短である、 該未符号化画素と同じ水平走査線上に位置する符号化済画素の画素値を、 該 未符号化画素の擬似画素値とするものであることを特徴とする画像符号化装 置。
4 . 請求の範囲第 1項記載の画像符号化装置において、
上記符号化手段は、 上記符号化済画素の画素値及び上記未符号化画素の擬 似画素値に基づいて被符号化画素に対する予測画素値を生成する予測値生成 器と、 上記被符号化画素の画素値と上記被符号化画素の予測画素値との差分 値を符号化し、 該符号化した差分値をプロック単位で画像符号化信号として 出力する符号化器とを有するものであることを特徴とする画像符号化装置。
5 . 請求の範囲第 1項または第 4項に記載の画像符号化装置において、 上記符号化手段は、 上記符号化済画素の画素値及び上記未符号化画素の擬 似画素値に基づいて、 上記被符号化画素の画素値を符号化するための符号語 を選択するものであることを特徴とする画像符号化装置。
6 . 請求の範囲第 1項または第 4項に記載の画像符号化装置であって、 上記符号化手段は、 上記符号化済画素の画素値及び上記未符号化画素の擬 似画素値に基づいて、 上記被符号化画素の画素値を算術符号化するための符 号コードに相当する確率テーブルを選択し、 上記被符号化画素に対する算術 符号化処理を、 該選択した確率テーブルに基づいて行うものであることを特 徴とする画像符号化装置。
7 . 請求の範囲第 1項記載の画像符号化装置において、
上記符号化済画素の画素値を復号化して局所複号化画素値を生成する局所 復号化手段を備え、
上記画素値置換手段を、 上記未符号化画素の画素値を、 上記符号化済画素 に対応する局所複号化画素値から所定の規則に基づいて得られる擬似画素値 と置換する構成とし、
上記符号化手段を、 上記符号化済画素の局所複号化画素値及び上記未符号 化画素の擬似画素値を参照して、 上記被符号化画素の画素値に対して非可逆 符号化処理を施す構成としたものであることを特徴とする画像符号化装置。
8 . 請求の範囲第 7項記載の画像符号化装置において、
上記符号化手段は、 上記符号化済画素の局所復号化画素値及び上記未符号 化画素の擬似画素値に基づいて被符号化画素に対する予測画素値を生成する 予測値生成器と、 上記被符号化画素の画素値と上記被符号化画素の予測画素 値との差分値を符号化し、 該符号化した差分値をプロック単位で画像符号化 信号として出力する符号化器とを有するものであることを特徴とする画像符 号化装置。
9 . 請求の範囲第 7項または第 8項に記載の画像符号化装置において、 上記符号化手段は、 上記符号化済画素の局所複号化画素値及び上記未符号 化画素の擬似画素値に基づいて、 上記被符号化画素の画素値を符号化するた めの符号語を選択するものであることを特徴とする画像符号化装置。
1 0 . 請求の範囲第 7項または第 8項に記載の画像符号化装置であって、 上記符号化手段は、 上記符号化済画素の局所符号化画素値及び上記未符号 化画素の擬似画素値に基づいて、 上記被符号化画素の画素値を算術符号化す るための符号コードに相当する確率テーブルを選択し、 上記被符号化画素に 対する算術符号化処理を、 該選択した確率テーブルに基づいて行うものであ ることを特徴とする画像符号化装置。
1 1 . 画像信号を構成する画素値を、 被符号化画素の周辺に位置する複数 の周辺画素の画素値に基づいて順次符号化する処理を、 所定数の画素からな るブロック毎に行って得られた画像符号化信号を、 該ブロック毎に復号化す る画像複号化装置であって、
被復号化画素の周辺に位置する複数の周辺画素のうちの未復号化画素の画 素値を、 該複数の周辺画素のうちの復号化済画素の画素値から所定の規則に 基づいて得られる擬似画素値と置換する画素値置換手段と、
上記各プロックに対応する複数の画素値からなる画像符号化信号を受け、 該各画素値を上記復号化済画素の画素値及び未復号化画素の擬似画素値に基 づいて順次復号化する複号化処理を、 上記ブロック単位で行って、 各プロッ クに対応した画像復号化信号を出力する複号化手段と、
上記各ブロックに対応した画像復号化信号を統合して、 走査線構造の画像 復号化信号に変換する逆プロック化手段を備え、
該走査線構造の画像復号化信号を 1画像表示画面に対応する画像再生信号 として出力することを特徴とする画像複号化装置。
1 2 . 請求の範囲第 1 1項記載の画像復号化装置において、
上記画素値置換手段は、 上記未復号化画素からの空間距離が最短である復 号化済画素の画素値を該未復号化画素の擬似画素値とするものであることを 特徴とする画像複号化装置。
1 3 . 請求の範囲第 1 1項記載の画像復号化装置において、
上記画素値置換手段は、 上記未複号化画素からの空間距離が最短である、 該未復 "^化画素と同じ水平走査線上に位置する複号化済画素の画素値を、 該 未複号化画素の擬似画素値とするものであることを特徴とする画像復号化装
1 4 . 請求の範囲第 1 1項記載の画像復号化装置において、
上記複号化手段は、
上記復号化済画素の画素値及び未複号化画素の擬似画素値に基づいて被復 号化画素に対する予測画素値を生成する予測値生成器と、
上記被複号化画素の画素値を復号化して復号化画素値を生成し、 該復号化 画素値に該被復号化画素の予測画素値を加算する復号化器とを有するもので あることを特徴とする画像復号化装置。
1 5 . 請求の範囲第 1 1項又は第 1 4項に記載の画像復号化装置において、 上記復号化手段は、 上記復号化済画素の画素値および上記未復号化画素の 擬似画素値に基づいて、 上記被復号化画素の画素値を復号化するための符号 語を選択するものであることを特徴とする画像複号化装置。
1 6 . 請求の範囲第 1 1項又は第 1 4項に記載の画像複号化装置において、 上記複号化手段は、 上記復号化済画素の画素値及び未複号化画素の擬似画 素値に基づいて、 上記被復号化画素の画素値を算術復号化するための符号コ 一ドに相当する確率テーブルを選択し、 上記被複号化画素に対する算術復号 化処理を、 該選択した確率テーブルに基づいて行うものであることを特徴と する画像複号化装置。
1 7 . 画像信号を構成する画素値を、 被符号化画素の周辺に位置する複数 の周辺画素の画素値に基づいて順次符号化する符号化処理を、 所定数の画素 からなるプロック毎に行う画像符号化方法であつて、
上記複数の周辺画素のうちの符号化済画素については、 その画素値を参照 画素値として設定し、 かつ上記複数の周辺画素のうちの未符号化画素につい ては、 該周辺画素のうちの符号化済画素の画素値から所定の規則に基づいて 得られる擬似画素値を参照画素値として設定し、
上記被符号化画素に対する複数の周辺画素に対して設定された参照画素値 に基づいて、 該被符号化画素の画素値を符号化して、 上記画像信号に対する 画像符号化信号を生成することを特徴とする画像符号化方法。
1 8 . 画像信号を構成する画素値を、 被符号化画素の周辺に位置する複数 の周辺画素の画素値に基づいて順次符号化する処理を、 所定数の画素からな るプロック毎に行って得られた画像符号化信号を、 該プロック毎に復号化す る画像複号化方法であって、
複数の周辺画素のうちの復号化済画素については、 その画素値を参照画素 値として設定し、 かつ上記複数の周辺画素のうちの未複号化画素については、 該周辺画素のうちの復号化済画素の画素値から所定の規則に基づいて得られ る擬似画素値を参照画素値として設定し、
被複号化画素に対する複数の周辺画素に対して設定された参照画素値に基 づいて、 該被復号化画素の画素値を復号化して、 上記画像符号化信号に対す る画像復号化信号を生成することを特徴とする画像複号化方法。
1 9 . 画像信号の符号化処理あるいは復号化処理を行うためのプログラム を格納したデータ記憶媒体であって、
上記プログラムは、 請求の範囲第 1 7項記載の画像符号化方法または第 1 8項記載の画像復号化方法による画像信号の処理をコンピュータに行わせる ための画像処理プログラムであることを特徴とするデータ記憶媒体。
PCT/JP1997/003917 1996-10-30 1997-10-29 Image encoder, image encoding method, image decoder, image decoding method, and data recording medium WO1998019461A1 (en)

Priority Applications (6)

Application Number Priority Date Filing Date Title
US09/077,378 US6173078B1 (en) 1996-10-30 1997-10-29 Image encoder, image encoding method, image decoder, image decoding method, and data recording medium
JP52029098A JP3176378B2 (ja) 1996-10-30 1997-10-29 画像復号化装置及び画像復号化方法
EP97909663A EP0873015B1 (en) 1996-10-30 1997-10-29 Image encoder, image encoding method, image decoder, image decoding method, and data recording medium
DE69709847T DE69709847T2 (de) 1996-10-30 1997-10-29 Verfahren und vorrichtung zur bildkodierung und -dekodierung sowie datenaufzeichnungsmedium
US09/867,610 US6487315B2 (en) 1996-10-30 2001-05-31 Image decoding apparatus using pixel values from at least three reference pixels
US10/238,600 US6748117B2 (en) 1996-10-30 2002-09-11 Image decoding apparatus using pixel values from at least three reference pixels

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP28785896 1996-10-30
JP8/287858 1996-10-30

Related Child Applications (3)

Application Number Title Priority Date Filing Date
US09/077,378 A-371-Of-International US6173078B1 (en) 1996-10-30 1997-10-29 Image encoder, image encoding method, image decoder, image decoding method, and data recording medium
US09/651,120 Continuation US6487314B1 (en) 1996-10-30 2000-08-30 Image decoding apparatus using pixel values from at least three references pixels
US09/651,120 Division US6487314B1 (en) 1996-10-30 2000-08-30 Image decoding apparatus using pixel values from at least three references pixels

Publications (1)

Publication Number Publication Date
WO1998019461A1 true WO1998019461A1 (en) 1998-05-07

Family

ID=17722684

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP1997/003917 WO1998019461A1 (en) 1996-10-30 1997-10-29 Image encoder, image encoding method, image decoder, image decoding method, and data recording medium

Country Status (12)

Country Link
US (4) US6173078B1 (ja)
EP (4) EP1075145B1 (ja)
JP (1) JP3176378B2 (ja)
KR (1) KR100287641B1 (ja)
CN (3) CN1165182C (ja)
DE (4) DE69718219D1 (ja)
ES (4) ES2190778T3 (ja)
ID (1) ID21363A (ja)
MX (1) MX9805292A (ja)
MY (1) MY124128A (ja)
TW (1) TW366669B (ja)
WO (1) WO1998019461A1 (ja)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013518456A (ja) * 2010-01-25 2013-05-20 トムソン ライセンシング 画像ブロックの符号化方法及び復号方法
CN103201598A (zh) * 2010-09-03 2013-07-10 洛斯阿拉莫斯国家安全股份有限公司 使用超声波使悬浮在流体中的微粒和流体流型可视化的装置和方法

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000115783A (ja) * 1998-10-06 2000-04-21 Canon Inc 復号化装置及び方法
GB2346470B (en) * 1999-02-05 2003-10-29 Advanced Risc Mach Ltd Bitmap font data storage within data processing systems
JP3885413B2 (ja) * 1999-06-08 2007-02-21 富士ゼロックス株式会社 画像符号化装置、画像復号装置および画像符号化復号装置ならびに方法
US6795367B1 (en) * 2000-05-16 2004-09-21 Micron Technology, Inc. Layout technique for address signal lines in decoders including stitched blocks
US7221483B2 (en) * 2000-09-05 2007-05-22 Ricoh Company, Ltd. Image encoding method and apparatus, image decoding method and apparatus, image processing apparatus, image formation apparatus, and computer-executable programs
JP4368575B2 (ja) 2002-04-19 2009-11-18 パナソニック株式会社 可変長復号化方法、可変長復号化装置およびプログラム
JP4090862B2 (ja) * 2002-04-26 2008-05-28 松下電器産業株式会社 可変長符号化方法および可変長復号化方法
CN101304536B (zh) 2002-11-20 2011-11-02 松下电器产业株式会社 动态图像编码方法及动态图像编码装置
JP4522199B2 (ja) * 2003-11-27 2010-08-11 キヤノン株式会社 画像符号化装置及び画像処理装置及びそれらの制御方法、並びにコンピュータプログラム及びコンピュータ可読記憶媒体
US7418146B2 (en) * 2004-02-10 2008-08-26 Sanyo Electric Co., Ltd. Image decoding apparatus
CN100426869C (zh) * 2006-07-21 2008-10-15 北京中星微电子有限公司 一种视频预测编码方法和解码方法
EP2071355B1 (en) 2007-12-13 2015-07-29 Swisscom AG System and method for determining a location area of a mobile user
US9721148B2 (en) 2007-12-31 2017-08-01 Applied Recognition Inc. Face detection and recognition
KR101517768B1 (ko) 2008-07-02 2015-05-06 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
US8295364B2 (en) * 2009-04-02 2012-10-23 Sony Corporation System and method of video data encoding with minimum baseband data transmission
KR101464538B1 (ko) * 2009-10-01 2014-11-25 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
WO2011042905A1 (en) * 2009-10-08 2011-04-14 Hd Medix Distributed video compression system
KR101798079B1 (ko) * 2010-05-10 2017-11-16 삼성전자주식회사 픽셀값의 차분을 이용하여 영상 프레임을 부호화하는 방법 및 이를 위한 장치
JP2014075631A (ja) * 2012-10-02 2014-04-24 Samsung Display Co Ltd 圧縮器、駆動装置および表示装置
KR101485463B1 (ko) * 2013-04-30 2015-01-26 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
KR101606855B1 (ko) * 2014-09-01 2016-03-28 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
KR101606856B1 (ko) * 2014-10-29 2016-03-28 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
KR101677277B1 (ko) * 2015-06-17 2016-11-17 삼성전자주식회사 영상의 부호화 방법 및 장치, 그 복호화 방법 및 장치
US20190364284A1 (en) * 2017-01-16 2019-11-28 Industry Academy Cooperation Foundation Of Sejong University Image encoding/decoding method and device
KR101989997B1 (ko) * 2018-06-27 2019-06-17 삼성전자주식회사 영상의 부호화 방법 및 장치, 및 컴퓨터로 판독 가능한 기록 매체

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS62122479A (ja) * 1985-11-22 1987-06-03 Canon Inc 画像伝送システム
JPH01205663A (ja) * 1988-02-12 1989-08-18 Nec Corp 画像信号符号化方式
JPH02105790A (ja) * 1988-10-14 1990-04-18 Sony Corp 画像信号の高能率符号化装置
JPH04219074A (ja) * 1990-08-31 1992-08-10 Toshiba Corp 画像符号化装置
JPH0678294A (ja) * 1992-08-21 1994-03-18 Sony Corp 符号化方法、符号化装置、および復号化装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5007102A (en) * 1986-03-20 1991-04-09 At&T Bell Laboratories Data compression using block list transform
CA1326899C (en) 1988-10-14 1994-02-08 Tetsujiro Kondo Highly efficient coding apparatus
JP2621747B2 (ja) * 1992-10-06 1997-06-18 富士ゼロックス株式会社 画像処理装置
US5883976A (en) * 1994-12-28 1999-03-16 Canon Kabushiki Kaisha Selectively utilizing multiple encoding methods
US5905819A (en) * 1996-02-05 1999-05-18 Eastman Kodak Company Method and apparatus for hiding one image or pattern within another
US5745603A (en) * 1996-08-26 1998-04-28 Hewlett-Packard Company Two dimensional context model obtained without a line buffer for arithmetic coding

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS62122479A (ja) * 1985-11-22 1987-06-03 Canon Inc 画像伝送システム
JPH01205663A (ja) * 1988-02-12 1989-08-18 Nec Corp 画像信号符号化方式
JPH02105790A (ja) * 1988-10-14 1990-04-18 Sony Corp 画像信号の高能率符号化装置
JPH04219074A (ja) * 1990-08-31 1992-08-10 Toshiba Corp 画像符号化装置
JPH0678294A (ja) * 1992-08-21 1994-03-18 Sony Corp 符号化方法、符号化装置、および復号化装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP0873015A4 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013518456A (ja) * 2010-01-25 2013-05-20 トムソン ライセンシング 画像ブロックの符号化方法及び復号方法
US9363514B2 (en) 2010-01-25 2016-06-07 Thomson Licensing Spatial prediction technique for video coding
CN103201598A (zh) * 2010-09-03 2013-07-10 洛斯阿拉莫斯国家安全股份有限公司 使用超声波使悬浮在流体中的微粒和流体流型可视化的装置和方法
CN103201598B (zh) * 2010-09-03 2017-05-24 洛斯阿拉莫斯国家安全股份有限公司 使用超声波使悬浮在流体中的微粒和流体流型可视化的装置和方法

Also Published As

Publication number Publication date
EP0873015A4 (en) 2000-08-02
EP1075144B1 (en) 2003-01-02
US20010024526A1 (en) 2001-09-27
EP1075144A1 (en) 2001-02-07
CN1122420C (zh) 2003-09-24
ES2190778T3 (es) 2003-08-16
EP0873015A1 (en) 1998-10-21
US20030007697A1 (en) 2003-01-09
CN1165182C (zh) 2004-09-01
CN1416278A (zh) 2003-05-07
EP0873015B1 (en) 2002-01-02
ES2187418T3 (es) 2003-06-16
DE69709847T2 (de) 2002-09-19
ES2170368T3 (es) 2002-08-01
CN1206535A (zh) 1999-01-27
US6487314B1 (en) 2002-11-26
CN1377190A (zh) 2002-10-30
US6173078B1 (en) 2001-01-09
DE69719100D1 (de) 2003-03-20
DE69719100T2 (de) 2003-07-17
TW366669B (en) 1999-08-11
EP1075145B1 (en) 2003-02-12
JP3176378B2 (ja) 2001-06-18
MX9805292A (es) 1998-10-31
KR100287641B1 (ko) 2003-01-24
DE69739867D1 (de) 2010-06-10
DE69709847D1 (de) 2002-02-28
CN1154369C (zh) 2004-06-16
ID21363A (id) 1999-05-27
ES2343885T3 (es) 2010-08-12
EP1271961B1 (en) 2010-04-28
EP1271961A3 (en) 2005-04-27
EP1075145A1 (en) 2001-02-07
KR19990076900A (ko) 1999-10-25
US6748117B2 (en) 2004-06-08
US6487315B2 (en) 2002-11-26
DE69718219D1 (de) 2003-02-06
EP1271961A2 (en) 2003-01-02
MY124128A (en) 2006-06-30

Similar Documents

Publication Publication Date Title
WO1998019461A1 (en) Image encoder, image encoding method, image decoder, image decoding method, and data recording medium
US5481554A (en) Data transmission apparatus for transmitting code data
JP4769717B2 (ja) 画像復号化方法
JP3762392B2 (ja) 動画像符号化方法及び装置
US7227900B2 (en) Differential encoding
TWI827606B (zh) 網格寫碼之量化係數寫碼
US20130272422A1 (en) System and method for encoding/decoding videos using edge-adaptive transform
JP2008278091A (ja) 動画像記録方法及びその装置
EP1322121A2 (en) Video encoder and decoder with improved motion detection precision
US20090317008A1 (en) Method and apparatus for encoding/decoding image using adaptive distribution adjustment of differential values
KR20050072491A (ko) 하이브리드 고해상도 dvd를 위한 2-층 인코딩 기술
EP0606856B1 (en) Digital video recording device with variable speed reproduction
CN113475067A (zh) 视频编解码方法及装置
KR100537123B1 (ko) 화상 신호 부호화 장치, 화상 신호 부호화 방법, 화상 신호 복호 장치, 화상 신호 복호 방법 및 기록매체
JP3735244B2 (ja) 画像符号化装置及び画像符号化方法
JP2003333600A (ja) 画像符号化方法および画像復号方法
CN112399180B (zh) 视频解码方法、装置及存储介质
JP3176601B1 (ja) 画像復号化装置
KR19980087387A (ko) 가변율 부호화 장치
JP2001157212A5 (ja)
MX2014004501A (es) Metodo de codificacion con prediccion de video de movimiento, dispositivo de codificacion con prediccion de video de movimiento, programa de codificacion con prediccion de video de movimiento, metodo para descifrar con prediccion de video de movimiento, dispositivo para descifrar con prediccion de video de movimiento, y programa para descifrar con prediccion de video de movimiento.
US20010043653A1 (en) Method and apparatus for image encoding method and appartus for image decoding and recording medium
JP3902698B2 (ja) 動画像の可逆圧縮符号化装置および可逆伸長復号化装置
JP2002027478A (ja) 画像情報符号化装置及び方法
JPH07107435A (ja) 光ディスク記録再生装置

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 97191534.2

Country of ref document: CN

AK Designated states

Kind code of ref document: A1

Designated state(s): CN ID JP KR MX SG US

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE

WWE Wipo information: entry into national phase

Ref document number: 1997909663

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: PA/a/1998/005292

Country of ref document: MX

Ref document number: 1019980705028

Country of ref document: KR

WWE Wipo information: entry into national phase

Ref document number: 09077378

Country of ref document: US

121 Ep: the epo has been informed by wipo that ep was designated in this application
WWP Wipo information: published in national office

Ref document number: 1997909663

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1019980705028

Country of ref document: KR

WWG Wipo information: grant in national office

Ref document number: 1019980705028

Country of ref document: KR

WWG Wipo information: grant in national office

Ref document number: 1997909663

Country of ref document: EP