JP3942611B2 - Decoding device and transmission system - Google Patents

Description

  The present invention relates to a device that decodes predictive-encoded (compression-encoded) moving image data, and a system that transmits the predictive-encoded (compressed-encoded) and decoding-related moving image data. In particular, the present invention relates to a function for improving a decrease in image quality due to a data error or loss on a transmission path or a propagation path.

  In recent years, devices for transmitting video such as video distribution systems such as video conferencing and VOD (video on demand) have been developed and spread. Against this background, international standards for moving picture coding methods have also been established. Recent international standards include ITU-T recommendation H.264. 261, H.H. 262, and H.H. H.263 and IS0 MPEG.

  In these international standards, pixel information is divided into several block units and encoded. This corresponds to a block of 8 pixels × 8 pixels and a macro block of 16 pixels × 16 pixels. Here, they are collectively referred to as a block. In addition, intra-frame encoding or inter-frame encoding is used when encoding pixel values. Intra-frame encoding is performed by orthogonally transforming pixel values for each block, whereas inter-frame encoding is performed by orthogonally transforming difference values using an image of the previous frame or the like as a reference image. A block coded by intraframe coding is called an intra block, and a block coded by interframe coding is called an inter block.

  Since coding by inter-frame coding often has a smaller code amount than intra-frame coding, many blocks are inter-frame coded and become inter blocks.

  A predetermined number of blocks subjected to orthogonal transformation are encoded together in units such as GOB (Group Of Blocks), GOP (Group Of Picture), slices, and video packets. Here, the coding unit is generically called a slice.

  An example of the encoded data is shown in FIG. The encoded data includes a slice header, a motion vector, a DC (direct current) value obtained by orthogonal transformation, and an AC (alternating current) value. For example, as in encoded data example 1 shown in FIG. 2A, each macroblock data following the slice header includes a motion vector, a DC value, and an AC value.

  In addition, a variable length code is often used for encoded data in order to improve the compression ratio. However, when such an encoding method is used, once an error occurs in the encoded data, at the time of decoding, errors after the occurrence of the error occur. Data cannot be decrypted correctly. Therefore, when adopting such an encoding method, it is normal to include a synchronization code in the encoded data. Even when decoding encoded data containing an error, the synchronization state is again restored after a certain interval has elapsed. The decryption process can be continued after recovery.

  Note that the encoded data example 2 illustrated in FIG. 2B is an example of encoded data in the case where motion vectors, DC values, and AC values are encoded together in all macroblocks included in one slice. is there. The encoded data 3 shown in FIG. 2 (c) has the same configuration as that of the encoded data example 2 shown in FIG. 2 (b), but is encoded when a synchronization code is inserted in the middle of the encoded data. It is an example of data.

  By the way, when the encoded data output by the encoding device is transmitted to the encoding device through the transmission path, there is a problem that a data error occurs during transmission. If a data error occurs in the encoded data, the decoding device cannot correctly decode the original data, so the influence of the error is large. In particular, when transmitting encoded data encoded by the inter-frame encoding method, once an image deteriorates due to an error, the next frame is decoded with reference to the image, so that the influence of the error is Propagates to the next and subsequent frames.

For this reason, the encoding device requires a reference image as a means for avoiding such error propagation, even if the code amount is large (although inter-frame coding is usually performed to reduce the code amount). There is also a method for forcibly performing intra-frame coding that is not performed, that is, a method for forcing an encoded block to be an intra block. The stomach down tiger block generated at this time is referred to as a forced intra block.

  Hereinafter, a conventional example will be described with reference to the drawings. Here, FIG. 3 shows a conventional example of a coding apparatus, and FIGS. 4 and 5 show a conventional example of a decoding apparatus.

  First, the operation of encoding apparatus 300 shown in FIG. 3 will be described. When frame data is input, the encoding unit 301 starts encoding in order for each block. Here, the encoding unit 301 performs intra-frame encoding for each block, or inter-frame encoding based on reference image frame data written in the reference memory 302.

  Also, the encoding unit 301 calculates a generated code amount in intra-frame encoding and a generated code amount in inter-frame encoding, and notifies the intra / inter determining unit 303 of the calculated generated code amount. The encoding unit 301 notifies the forced intra determination unit 304 of the frame serial number and the block serial number.

  The compulsory intra determination unit 304 periodically determines a block to be intraframe-encoded based on the block serial number and the frame serial number so that the entire frame is always intraframe-encoded during encoding of several frames. Instruct the intra / inter determination unit 303 to that effect.

  When there is no instruction from the forced intra determination unit 304, the intra / inter determination unit 303 selects an encoding method based on the notified generated code amount and performs intra-frame encoding from the forced intra determination unit 304. Is notified, intra-frame coding is selected. The intra / inter determination unit 303 notifies the encoding unit 301 of the selected encoding method in any case.

  The encoding unit 301 determines the notified encoding method as a block encoding method to be encoded. As shown in FIG. 2, the encoding unit 301 encodes the header, the motion vector information of each block, the DC value of each block, and the AC value of each block as shown in FIG. Output as data.

  Note that the decoding unit 305 decodes the encoded data output by the encoding unit 301 and writes it in the reference memory 302. The encoding apparatus 300 repeats the encoding of the next frame every time the encoding procedure of one frame data is completed by such a procedure.

  Next, the operation of the decoding device 400 shown in FIGS. 4 and 5 will be described. FIG. 5 illustrates a configuration example of the image configuration unit 402. As illustrated in FIG. 4, the decoding device 400 includes a decoding unit 401, an image configuration unit 402, a motion vector interpolation unit 403, and a reference memory 404.

  The decoding unit 401 decodes the input encoded data based on the grammar (encoding method). That is, the encoding unit 401 obtains reference image data from the reference memory 404 and decodes the encoded data that has been inter-frame encoded, and decodes the encoded data that has been intra-coded as it is. To do.

  The decoding unit 401 outputs the decoded image data to the image configuration unit 402, and outputs the motion vector to the motion vector interpolation unit 403. If the decoding unit 401 detects an error in the encoded data at the time of decoding, the decoding unit 401 notifies the image forming unit 402 to that effect as an error notification signal.

  The motion vector interpolating unit 403 obtains a motion vector from the decoding unit 401, but obtains a motion vector by performing interpolation for a block in which no grammatical motion vector exists or a block in which a motion vector could not be decoded. This is notified to the unit 402 as a motion vector.

  When there is no error notification from the decoding unit 401, the image construction unit 402 switches the input of the changeover switch 501 (FIG. 5) to the decoding unit 401 side, and the image data given from the decoding unit 401 is converted into the frame data construction unit 502. (Fig. 5) On the other hand, when there is an error notification from the decoding unit 401, the image configuration unit 402 switches the input of the changeover switch 501 (FIG. 5) to the previous image processing unit 503 (FIG. 5) and outputs from the previous image processing unit 503. The interpolated image is transferred to the frame data construction unit 502 (FIG. 5).

  Here, the previous image processing unit 503 forms an interpolation image from the motion vector information given from the motion vector interpolation unit 401 (FIG. 4) and the reference image read from the reference memory 404 (FIG. 4).

  The frame data construction unit 502 (FIG. 5) configures image data for one frame when the image data is prepared for one frame, and outputs this as frame data. As described above, when one frame of image data (frame data) is configured by the frame data construction unit 502, the image configuration unit 402 writes the image data in the reference memory 404 and outputs it as frame data. Thus, the reference memory 404 always holds the latest image data (frame data) configured by the image configuration unit 402.

Note that there are Patent Document 1 and Patent Document 2 as documents that refer to a repair method when an error occurs.
Japanese Patent Laid-Open No. 9-247681 JP-A-9-200743

  In general, even if the inter block is wrong, the difference information of each pixel is small and the error itself is small. However, if the intra block is wrong, the influence is large. turn into.

  For this reason, conventionally, an intra-block error concealment method has been considered. As a typical intra block error concealment method, a method of concealing an error by displaying a previous image instead of an error occurrence part, and concealing an error by displaying the previous image indicated by a motion vector using motion information And a method of concealing an error by displaying an image decoded from only a DC (direct current) component.

  However, the method of displaying the previous image can effectively conceal the error when the correlation with the previous image is large, while the image is distorted when the correlation with the previous image is small. Further, in the case of a method of displaying an image decoded from only a DC (direct current) component, it is not necessary to display a remarkably distorted image. On the other hand, when an error occurs in a portion having a large correlation with the previous image, It is often said that displaying the previous image is a visually superior error concealment method.

  Incidentally, when an intra block occurs, it can be considered that the amount of generated code is smaller in intra-frame coding than in inter-frame coding and the case of forced intra frame and forced intra block, In the former case, there is little correlation with the previous image, and in the latter case, there is often a correlation with the previous image.

  For this reason, when a moving image that does not match the error concealment method to be used becomes a decoding target, there is a problem that image quality deterioration is easily noticeable.

The present invention has been made to solve the above technical problems, an object of proposals noticeable difficulty had decrypted device and transmission system degradation in image quality compared to the conventional.

  In the first aspect of the present invention, a decoding apparatus including decoding means for decoding and outputting encoded data that has been predictively encoded includes the following means.

That is, (1) when an error is detected for an intra-encoded block during the decoding operation of the decoding means, the encoded data is forcibly intra-encoded from the encoding device. Check whether the signal to be notified is input, and if the signal is input, determine the error concealment process referring to the previous image. If the signal is not input, it is decoded. A compulsory intra signal receiving unit that determines an error concealment process based on a DC value; and (2) if the decoding unit performs a decoding operation normally, an original image is constructed based on the decoded data, If an error is detected for intra-coded blocks in the decoding operation of the means to construct an image of the corresponding portion based on concealment processing method was from the forced intra signal receiving means of the notification So that and an image constructing unit.

  Here, it is highly likely that an intra-encoded image that has not been forcibly encoded intra-encoded has a low correlation with the previous image in the determination of the encoding efficiency, so the error is concealed by the previous image. The subjective image quality can be expected to be improved by concealing the error with the DC value rather than the case of doing so.

In the second present invention, in order to suppress error propagation, a compulsory intra determination unit that determines execution of a compulsory intra coding process as appropriate at the time of encoding, and a compulsory intra code by the compulsory intra determination unit A compulsory intra notification means for transmitting a compulsory intra signal for notifying that the corresponding encoded data is encoded data that has been compulsorily intra-coded when the determination of the encoding process is notified to the decoding device; And a decoding apparatus according to the first aspect of the present invention constitute a transmission system. As a result, it is possible to perform a decoding operation by adding information on whether intra-coded data is forcibly intra-coded or intra-coded from the viewpoint of coding efficiency. Possible transmission systems can be made possible.

  According to the first aspect of the present invention, the error concealment method is selected based on whether or not the data to be decoded is forcibly intra-coded, and the intra-encoding is forcibly performed. It is highly likely that an intra-coded image that is not is determined to have a low correlation with the previous image in the determination of the coding efficiency, so the error is concealed by the decoded DC value. This makes it possible to achieve a subjective improvement in image quality compared to concealing errors.

According to the second aspect of the present invention, in order to suppress error propagation, a compulsory intra determination unit that determines execution of a compulsory intra coding process as needed during encoding, Compulsory intra notification means for transmitting a compulsory intra signal for notifying that the corresponding encoded data is encoded data that has been compulsorily intra-encoded to the decoding apparatus when the encoding process decision is notified And the decoding apparatus according to the first aspect of the present invention constitutes a transmission system, so that intra-coded data is forcibly intra-coded or encoded. From the viewpoint of efficiency, it is possible to make a transmission system capable of selecting an error concealment method by adding information indicating whether it is intra-encoded.

(A) Moving Image Transmission System Hereinafter, embodiments of the moving image transmission system according to the present invention will be described. In addition, each of these video transmission systems uses one or a plurality of video encoding devices and video decoding devices as transmission paths (when all paths are propagation paths and some paths are propagation paths). In addition to the case where real-time transmission of moving images is realized, the transmission of moving images is performed through recording and reproduction on various storage media (for example, optical discs). It can also be applied to non-real-time systems that are realized.

  However, in the following description, a case of a moving image transmission system including one moving image encoding device and one moving image decoding device, which is the most basic connection form, will be described.

(B) First Embodiment In the first embodiment, a moving image transmission system including the encoding device 300 shown in FIG. 3 and the decoding device 100 shown in FIG. 1 will be described. That is, a moving image transmission system configured by a combination of an existing encoding device 300 and a new decoding device 100 will be described.

  Note that the decoding apparatus 100 used in this embodiment is a case where an error is detected from an intra block, and the DC value decoded at the error detection location is greatly different from the DC values of the surrounding (including adjacent) blocks. In this case, a feature is that a function for concealing an error based on the previous frame image is mounted instead of concealing an error by a DC value.

(B-1) Configuration of Encoding Device 300 and Decoding Device 100 (B-1-1) Configuration of Encoding Device 300 As described above, the encoding device 300 includes the encoding unit 301, the reference memory 302, The intra / inter determination unit 303, the forced intra determination unit 304, and the decoding unit 305 are configured. Here, intra and inter represent intra-frame coding and inter-frame coding, respectively.

  The encoding unit 301 is a means for encoding the input frame data for each block, and depending on the format and encoding mode adopted as the encoding method, encoded data as represented in FIG. Output. The encoding unit 301 switches the encoding method in response to the notification from the intra / inter determination unit 303 at the time of encoding, and information on whether the frame data has been encoded by intraframe encoding or interframe encoding. Add to the encoded data and output.

  The encoding unit 301 has a function of notifying the intra / inter determination unit 303 of the generated code amount at the time of intra-frame encoding and the generated code amount at the time of inter-frame encoding. The encoding unit 301 performs encoding by referring to the frame data held in the reference memory 302 at the time of inter-frame encoding. The encoding unit 301 gives information necessary for determining a forced intra frame (block) to the forced intra determination unit 304, such as what number block in the frame the currently encoded block is.

  The decoding unit 305 is means for decoding the encoded data encoded by the encoding unit 301 and writing the decoded frame data into the reference memory 302. The reference memory 302 has a function of holding the frame data obtained from the decoding unit 305 and a function of outputting the frame data to the encoding unit 301.

  The compulsory intra determination unit 304 is a unit that determines compulsorily performing intraframe encoding based on information obtained from the encoding unit 301. For example, the forced intra determination unit 304 determines the serial number of the frame notified by the encoding unit 301, Then, a compulsory intra designation signal for instructing the compulsory encoding by intra-frame encoding is periodically output to the intra / inter determination unit 303 from the serial number of the block.

  The intra / inter determination unit 303 normally determines which of the intra-frame coding and the inter-frame coding is used based on the generated code amount information provided from the coding unit 301. However, when the intra / inter determination unit 303 receives a forced intra designation signal from the forced intra determination unit 304, the intra / inter determination unit 303 determines encoding by intraframe encoding. When the encoding method is determined, the intra / inter determination unit 303 notifies the encoding unit 301 of the encoding method.

(B-1-2) Configuration of Decoding Device 100 The decoding device 100 includes a decoding unit 101, an image configuration unit 102, a DC determination unit 103, and a reference memory 104.

  The decoding unit 101 is means for decoding the received encoded data and decoding it into frame data. The decryption unit 101 includes three functional units. The first function is to output the motion vector, DC component, and AC component obtained at the time of decoding. The second function is to detect an error contained in a video packet from the mismatch of codewords and the presence or absence of inconsistencies such as decoding a motion vector referring to the outside of an image frame. The third function is to notify the DC determination unit 103 of an error detection signal indicating a block including a detected error.

  The DC determination unit 103 receives and holds the DC component of the decoded image data from the decoding unit 101, and then has a predetermined range with respect to the predicted DC component value obtained from the DC component of the surrounding (including adjacent) block. It is determined whether it is within. Here, the predicted DC component value includes, for example, an average value of DC component values of adjacent blocks located at the upper, lower, left, right, upper left, lower left, upper right, and lower right of the decoded block, An average value or the like obtained by multiplying the DC component value of the adjacent block by a predetermined weight is used. Further, the predetermined range is determined by, for example, multiplying the variance value with respect to the predicted average of the DC component values by a coefficient.

  Upon receiving the error detection signal from the decoding unit 101, the DC determination unit 103 determines whether or not the DC component is within a predetermined range. If it is determined that the DC component is within the predetermined range, an error is caused by the DC component. Is notified to the image construction unit 102 as an error concealment method notification signal. On the other hand, when it is determined that the DC component does not fall within the predetermined range, the DC determination unit 103 notifies the image configuration unit 102 as an error concealment method notification signal that an error should be concealed with the previous frame image. Thus, the DC determination unit 103 prevents the image from being disturbed and uncomfortable due to the corresponding part being always concealed by the DC image.

  Note that when the DC component is not provided from the decoding unit 101 (when decoding is not possible), the DC determination unit 103 notifies the image configuration unit 102 that the error should be concealed with the previous frame image. Notify as a notification signal.

  When the error concealment method is not notified from the DC determination unit 103 (that is, in a normal state), the image configuration unit 102 decodes image information, DC component, AC component, and motion vector information provided from the decoding unit 101. Based on the above, an image is constructed based on an image data encoding method.

  On the other hand, when the error concealment method is notified from the DC determination unit 103, the image construction unit 102 constructs an image based on the error concealment method. For example, when a method using a DC component is notified as an error concealment method, the image construction unit 102 constructs an image block of an error part using only the DC component. In addition, when the image construction unit 102 is notified of the method using the previous frame image as the error concealment method, the image part of the error part is the image part specified by the motion vector in the frame data of the reference frame read from the reference memory 302 Configure the block. It should be noted that the previous frame image of the block at the same position is displayed without using a motion vector for a block for which motion vector information has not been obtained while receiving notification that the previous frame image is to be displayed.

  FIG. 6 shows an internal block configuration of the image configuration unit 102. The image construction unit 102 includes a previous image processing unit 601, an AC image construction unit 602, a DC image construction unit 603, an addition unit 604, an error concealment method control unit 605, a changeover switch 606, and a frame data construction unit 607. It consists of.

  The previous image processing unit 601 forms a concealed image from the image data of the previous frame read from the reference memory 104 and the motion vector information given from the decoding unit 101, and transmits the concealed image to the changeover switch 606.

  The AC image construction unit 602 composes an image based on the AC information given from the decoding unit 101 and transmits the image to the addition unit 604.

  The DC image construction unit 603 composes an image based on the DC information given from the decoding unit 101 and transmits the image to the addition unit 604 and the changeover switch 606.

  The error concealment method control unit 605 receives the error concealment method notification signal from the DC determination unit 103 and controls the changeover switch 606 and the previous image processing unit 601. Note that the error concealment method control unit 605 reduces the processing load and power consumption by not operating the previous image processing unit 601 when the error concealment method using the previous image is not adopted.

  The changeover switch 606 is a means for receiving a block image configured according to each error concealment method and transmitting one of them to the frame data construction unit 607. Here, the changeover switch 606 performs switching of the input image according to a signal given from the error concealment method control unit 605. In the normal state, the changeover switch 606 transmits the block image output from the addition unit 604 to the frame data construction unit 607.

  The frame data construction unit 607 receives the block image data from the changeover switch 606 and constructs the frame image data. The frame data construction unit 607 writes the constructed frame image data in the reference memory 104 and outputs it to the subsequent processing circuit as frame data.

(B-2) Moving Image Data Transmission Operation and Error Concealment Operation First, the operation of the encoding apparatus 300 will be described. When the frame data is input to the encoding unit 310, the encoding unit 301 starts encoding the frame data in order of each block. The encoding unit 301 performs encoding by either intra-frame encoding or inter-frame encoding for each block. Note that the frame data of the reference image held in the reference memory 302 is used at the time of interframe coding. Thereafter, the encoding unit 301 calculates a generated code amount generated by intraframe encoding and a generated code amount generated by interframe encoding, and notifies each value to the intra / inter determination unit 303. The encoding unit 301 notifies the forced intra determination unit 304 of the frame serial number and the block serial number.

  The forced intra determination unit 304 periodically determines a block to be intra-frame-encoded based on the block serial number and the frame serial number so that the entire frame is always intra-frame encoded during encoding of several frames. This is instructed to the intra / inter determination unit 303.

  When there is no instruction from the forced intra determination unit 304, the intra / inter determination unit 303 selects an encoding method based on the notified generated code amount and performs intra-frame encoding from the forced intra determination unit 304. When the instruction is notified, it is selected to perform intra-frame encoding. The intra / inter determination unit 303 notifies the encoding unit 301 of the selected encoding method in any case.

  The encoding unit 301 determines the notified encoding method as a block encoding method to be encoded. As shown in FIG. 2, the encoding unit 301 encodes the header, the motion vector information of each block, the DC value of each block, and the AC value of each block as shown in FIG. Data is output to the subsequent circuit (transmission path or propagation path).

  Note that the decoding unit 305 decodes the encoded data output by the encoding unit 301 and writes it in the reference memory 302. The encoding apparatus 300 repeats the encoding of the next frame every time the encoding procedure of one frame data is completed by such a procedure.

  Next, the operation in the decoding device 100 will be described. When encoded data received by the decoding apparatus 100 from the transmission path or propagation path is input to the decoding unit 101, the decoding unit 101 sequentially outputs DC information (pixel data based on DC values), AC, for each block. Information (pixel data based on AC values) and motion vector information are output to the image construction unit 102. At this time, the decoding unit 101 also notifies the DC determination unit 103 of the DC value. The DC determination unit 103 holds the DC value of each block notified from the decoding unit 101.

  The decoding unit 101 obtains and decodes reference frame data from the reference memory 104 when the block to be decoded is inter-frame encoded at the time of decoding, and the block to be decoded is intra-frame encoded. If so, the block data itself is decoded.

  When an error is detected during decoding and detected in an intra block, the decoding unit 101 outputs the data decoded so far as it is, and notifies the DC determination unit 103 of the intra block that cannot be decoded. To do.

  Upon receiving the error notification, the DC determination unit 103 compares the DC value of the intra block that has not been correctly decoded with a predetermined reference value (for example, the average value of neighboring (including adjacent) blocks), and the DC value It is checked whether or not is greatly deviated from a predetermined reference value.

  If it is determined that the DC value is not within the predetermined range with respect to the predetermined reference value as a determination result here, the DC determination unit 103 is not allowed to construct an image that conceals the error. It is determined that the value is appropriate, and notifies the image construction unit 102 to select an error concealment method for displaying the image of the previous frame.

  On the other hand, when it is determined that the DC value is within a predetermined range with respect to the predetermined reference value, the DC determination unit 103 determines that the DC value is an appropriate value for constructing an image that conceals the error part. And notifies the image composition unit 102 to select an error concealment method for displaying an image constructed based on the DC value.

  In addition, when the DC value cannot be decoded in the decoding unit 101 and the DC determination unit 103 has not notified the DC value itself, the DC determination unit 103 conceals an error with the image of the previous frame. Notify you to select.

  These error concealment methods are given to the error concealment method control unit 605 of the image construction unit 102 as an error concealment method notification signal.

  When the notified error concealment method is based on a DC value, the error concealment method control unit 605 switches the input of the changeover switch 606 to the DC image configuration unit 603 side, and the image configured by the DC image configuration unit 603 is framed. The data construction unit 607 is informed. Thereby, the error part of frame data is concealed by the block image by DC value.

  In addition, when the notified error concealment method is based on the previous frame image, the error concealment method control unit 605 switches the input of the changeover switch 606 to the previous image processing unit 601 side and stores the previous information stored in the reference memory 104. A block image decoded from the frame image and the motion vector is transmitted to the frame data construction unit 607. Thereby, the error portion of the frame data is concealed by the block image obtained from the previous frame image.

  Note that, when the error concealment method is not notified from the DC determination unit 103, the error concealment method control unit 605 (FIG. 6) configures a block image using image data decoded from the DC value and AC value given from the decoding unit 101. The changeover switch 606 is controlled to do so. That is, the image construction unit 102 uses the changeover switch 606 to reconstruct the block images constructed in the AC image construction unit 602 and the DC image construction unit 603 shown in FIG. select. As a result, the block image composed of the decoded data is incorporated into a part of the frame data.

  The frame data construction unit 607 constructs all the block images constituting the frame image by executing the above procedure for each block. When the decoding process is completed for all areas of the frame image, the frame data construction unit 607 writes the decoded frame data into the reference memory 104 and outputs it to the subsequent circuit as frame data.

  Thus, the decoding process for one frame in the decoding apparatus 100 ends. After that, the decoding apparatus 100 repeats the decoding procedure of the encoded data in the same procedure to form the frame data of the next frame.

(B-3) Effect of First Embodiment As described above, when the decoding apparatus 100 proposed in the present embodiment is used, even when an error is detected during decoding of an intra block, decoding is performed at the error detection portion. When the DC value is significantly different from the DC value of the surrounding block, the error is not concealed using that value, and the block image constructed from the previous frame image is used for concealment. Images can be made difficult to output.

  Therefore, by combining the decoding apparatus 100 and the encoding apparatus 300, it is possible to realize a moving image transmission system that can be expected to improve subjective evaluation of image quality as compared with the conventional art.

(C) Second Embodiment In the second embodiment, a moving image transmission system including the encoding device 300 shown in FIG. 3 and the decoding device 700 shown in FIG. 7 will be described. That is, a moving image transmission system configured by a combination of an existing encoding device 300 and a new decoding device 700 will be described.

  It should be noted that the decoding apparatus 700 used in this embodiment example, when an error is detected from an intra block, is different from the DC value decoded from the error detection location and the DC value of surrounding (including adjacent) blocks. This is characterized in that a more appropriate concealment method can be selected by taking into account the magnitude of the motion vector decoded from the same location in switching the concealment method according to the above.

(C-1) Configuration of Encoding Device 300 and Decoding Device 700 The configuration of the encoding device 300 is the same as that of the first embodiment. Therefore, only the configuration of the decoding device 700 will be described below.

  The decoding apparatus 700 includes a decoding unit 701, an image configuration unit 702, a DC determination unit 703, a reference memory 704, a motion vector interpolation unit 705, a motion vector magnitude determination unit 706, and a DC / motion vector condition. And a determination unit 707.

  Among these, the decoding unit 701, the image configuration unit 702, the DC determination unit 703, and the reference memory 704 correspond to the decoding unit 101, the image configuration unit 102, the DC determination unit 103, and the reference memory 104 in the first embodiment. However, it is configured to realize equivalent functions.

  Therefore, the motion vector interpolation unit 705, the motion vector magnitude determination unit 706, and the DC / motion vector condition determination unit 707 are components unique to the present embodiment. However, the image construction unit 702 is different from the first embodiment in that interpolation motion vector information is added as an input to the previous image processing unit 801 as shown in FIG.

  Hereinafter, the functional configuration of the decoding device 700 will be described focusing on differences from the decoding device 100 in the first embodiment.

  When there is a block from which motion vector information cannot be obtained in the decoding unit 701, the motion vector interpolation unit 705 is a block that is spatially adjacent (including the periphery) to the block and has already been obtained. This means is provided for generating a corresponding motion vector (interpolated motion vector) from the existing one.

When the motion vector interpolation unit 705 receives an error detection notification for an intra block from the decoding unit 701, the motion vector received from the decoding unit 701 (when a motion vector is given from the decoding unit 701) or generated internally The interpolated motion vector (when no motion vector is given from the decoding unit 701) is notified to the motion vector magnitude determination unit 706 .

  Note that the motion vector interpolation unit 705 receives the error detection in the intra block from the decoding unit 701, and when the motion vector cannot be obtained from the decoding block, the interpolation motion vector. Is output to the image construction unit 702.

  The motion vector magnitude determining unit 706 compares the magnitude of the motion vector of the decoded block with a predetermined reference value (range), and the magnitude of the motion vector is significantly different from the predetermined reference value (range). It is a means to check whether there is any. Here, the predetermined reference value (range) is a range from “0 to the block size”, “from 0 to the macroblock range”, “0 to the block size. The range up to the value multiplied by the coefficient is considered.

  When it is determined that the size of the motion vector is not within a predetermined range with respect to a predetermined reference value (that is, when it is determined that the motion is large), the motion vector size determination unit 706 It is determined that the correlation with the frame is small, and notifies the DC / motion vector condition determination unit 707 to select an error concealment method for displaying an image configured based on the DC value.

  On the other hand, the motion vector magnitude determination unit 706 determines that the correlation between the block and the previous frame is large when it is determined that the magnitude of the motion vector is within a predetermined range with respect to a predetermined reference value. Then, the DC / motion vector condition determination unit 707 is notified to select an error concealment method for displaying the image of the previous frame.

  The DC / motion vector condition determination unit 707 receives notification of the error concealment method from each of the DC determination unit 703 and the motion vector magnitude determination unit 706. When the DC / motion vector condition determination unit 707 obtains a determination result indicating that the error concealment method based on the previous frame image should be selected from the motion vector magnitude determination unit 706, the determination is performed as a DC determination. The determination result of the unit 703 is prioritized, and the error concealment method using the previous frame image is set as the determination result. On the other hand, when the DC / motion vector condition determination unit 707 obtains a determination result from the motion vector magnitude determination unit 706 that the error concealment method based on the DC value should be selected because it is outside the predetermined range, the DC determination unit 703 The determination result is defined as the determination result.

  That is, the DC / motion vector condition determination unit 707 instructs to select an error concealment method based on the previous frame image for a block having a small motion vector size, and is decoded even if the motion vector size is large. For a block having a large difference between the DC value and the peripheral block, an instruction is given to select an error concealment method based on the previous frame image, and although the magnitude of the motion vector is large, the decoded DC value and the DC value of the peripheral block are For a block with a small difference, an instruction is given to select an error concealment method based on a DC value.

  The reason for this condition determination is that a portion where the degree of discontinuity of the DC value is large is easily noticeable due to human visual characteristics.

(C-2) Moving Image Data Transmission Operation and Error Concealing Operation Next, moving image data transmission operation and error concealing operation based on the moving image transmission system according to the present embodiment will be described. However, the configuration of the encoding apparatus 300 on the moving image transmission side is the same as that of the first embodiment, and the operation content is also the same, so the description thereof is omitted here. Hereinafter, an operation of the decoding apparatus 700 that performs a decoding process on encoded data received via a transmission path or a propagation path, particularly an error concealment operation will be described.

  When the encoded data received from the transmission path or the propagation path is input to the decoding unit 701, DC information (pixel data based on DC values), AC information (pixel data based on AC values), The motion vector information is decoded and output to the image construction unit 702. Further, the decoding unit 701 transmits the motion vector information of each block to the motion vector interpolation unit 705. The motion vector interpolation unit 705 holds the motion vector information of each block notified from the decoding unit 701.

  When an error is detected during decoding and is detected in an intra block, the decoding unit 701 outputs the data that has been decoded so far as it is, and outputs the intra block that cannot be decoded to the motion vector interpolation unit 705 and Notify the DC determination unit 703.

  When the motion vector interpolation unit 705 receives the error detection notification in the intra block, the motion vector decoded for the corresponding intra block (when the motion vector can be decoded by the decoding unit 701) or the average value of the motion vectors of adjacent blocks Interpolated motion vectors generated from the above (when the motion vector could not be decoded by the decoding unit 701) are held as motion vector information for the corresponding block, and are also stored in the image configuration unit 702 and the motion vector magnitude determination unit 706. Notice.

  When a motion vector (interpolated motion vector) is notified from the motion vector interpolating unit 705, the motion vector size determining unit 706 compares the motion vector size with the allowable range of the motion vector size, and the motion vector size is determined. Check if it is within the specified tolerance.

  As a result of the determination here, when it is determined that the size of the motion vector is within the allowable range, the motion vector size determination unit 706 determines that the motion is small and the correlation with the previous frame is high, and the The DC / motion vector condition determination unit 707 is notified to select an error concealment method using an image.

  On the other hand, when it is determined that the size of the motion vector is outside the allowable range, the motion vector size determination unit 706 determines that the correlation with the previous frame is low because the motion is large, and is configured based on the DC value. The DC / motion vector condition determination unit 707 is notified to select an error concealment method for displaying the displayed image.

  The DC determination unit 703 performs the same operation as the DC determination unit 103 described in the first embodiment. That is, the DC determination unit 703 compares the DC value of the intra block that has not been correctly decoded with a predetermined reference value (for example, the average value of neighboring (including adjacent) blocks), and the DC value is determined to be the predetermined reference value. Check if it is not significantly different from the value.

  When it is determined that the DC value is not within the predetermined range with respect to the predetermined reference value, the DC determination unit 703 is an inappropriate value for constructing an image that conceals an error. And the DC / motion vector condition determining unit 707 is notified to select an error concealment method for displaying the image of the previous frame.

  On the other hand, when it is determined that the DC value is within a predetermined range with respect to the predetermined reference value, the DC determination unit 703 determines that the DC value is an appropriate value for constructing an image that conceals the error part. And the DC / motion vector condition determining unit 707 is notified to select an error concealment method for displaying an image configured based on the DC value.

  If the decoding unit 701 cannot decode the DC value and the DC determination unit 703 does not notify the DC value itself, the DC determination unit 703 uses an error concealment method that conceals an error with the image of the previous frame. Notify you to choose.

  The DC / motion vector condition determination unit 707 receives notification of the error concealment method from each of the DC determination unit 703 and the motion vector magnitude determination unit 706, and selects an error concealment method based on the previous frame image from the motion vector magnitude determination unit 706. When the determination result that should be obtained is obtained, the determination is prioritized over the determination result of the DC determination unit 703, and the determination result that the error concealment method based on the DC value should be selected from the motion vector magnitude determination unit 706 is obtained. In this case, the determination result of the DC determination unit 703 operates so as to give priority.

  That is, the DC / motion vector condition determination unit 707 selects the error concealment method based on the previous frame image for a block with a small motion vector size, so that the decoded DC value and the motion vector size are large. For blocks with a large difference from neighboring blocks, the error concealment method based on the previous frame image is selected, so that the difference between the decoded DC value and the neighboring blocks is small, depending on the DC value. The image construction unit 702 is notified to select an error concealment method.

  The image construction unit 702 receives the error concealment method notification signal by the error concealment method control unit 805 as in the case of the image construction unit 102 of the first embodiment, and switches to the previous image processing unit 801 according to the instruction content. The switch 806 is switched and controlled. However, when the motion vector information is not obtained from the decoding unit 701, the previous image processing unit 801 uses the interpolation motion vector generated by the motion vector interpolation unit 705 instead.

  Operations such as a frame composition operation, a reference image write operation to the reference memory 704, and frame data output in the image construction unit 702 are the same as those in the first embodiment.

(C-3) Effects of Second Embodiment As described above, if the decoding apparatus 700 proposed in this embodiment is used, the following effects can be obtained in addition to the same effects as those of the first embodiment. it can. That is, in the present embodiment, for an intra block considered to have a small motion vector and a high correlation with the image of the previous frame, priority is given to the error concealment method that displays the image of the previous frame as its error concealment method. Therefore, the possibility that an image that impairs the subjectivity is output can be effectively avoided, and the subjective evaluation of the image quality when the intra block is erroneous can be improved.

(D) Third Embodiment In the third embodiment, a moving image transmission system including the encoding device 300 shown in FIG. 3 and the decoding device 900 shown in FIG. 9 will be described. That is, a moving image transmission system configured by a combination of an existing encoding device 300 and a new decoding device 900 will be described.

  Note that the decoding apparatus 700 used in this embodiment example further improves the concealment method by taking into account the amount of motion vector change in addition to the concealment method switching determination criterion used in the second embodiment example. It is characterized in that can be selected.

(D-1) Configuration of Encoding Device 300 and Decoding Device 900 The configuration of the encoding device 300 is the same as that of the first and second exemplary embodiments. Therefore, only the configuration of the decoding apparatus 900 will be described below.

  The decoding apparatus 900 includes a decoding unit 901, an image configuration unit 902, a DC determination unit 903, a reference memory 904, a motion vector interpolation unit 905, a motion vector magnitude determination unit 906, and a DC / motion vector condition. The determination unit 907 includes a motion vector change amount determination unit 908 and a motion vector condition determination unit 909.

  Among these, the differences from the second embodiment are two points: a motion vector change amount determination unit 908 and a motion vector condition determination unit 909. Therefore, the other parts are basically the same as those in the second embodiment.

  The motion vector change amount determination unit 908 is a unit that receives the notification of the motion vector information of each block from the motion vector interpolation unit 905 and holds this. When the motion vector change amount determination unit 908 receives a notification from the motion vector interpolation unit 905 that the intra block has not been correctly decoded, the motion vector change amount determination unit 908 uses the difference between the stored motion vector of the previous frame and the motion vector of the current frame. Then, the amount of change generated in the motion vector of the block to be decoded is calculated.

  The motion vector change amount determination unit 908 executes a function of selecting an error concealment method based on whether or not the change amount of the motion vector is within a predetermined range and notifying the motion vector condition determination unit 909. Here, the predetermined range is “with a range from 0 to the block size”, “with a size from 0 to the macroblock”, “0 to the block size multiplied by a predetermined coefficient. The range is up to the value ".

  The motion vector change amount determination unit 908 selects the error concealment method based on the image of the previous frame when the change amount of the motion vector is within a predetermined range (that is, when the change in motion is small). The vector condition determination unit 909 is notified.

  On the other hand, if the change amount of the motion vector is not within the predetermined range (that is, if the change of motion is large), the motion vector change amount determination unit 908 selects error concealment by the image of the previous frame. It is determined that the error concealment effect is reduced, and notifies the motion vector condition determination unit 909 to select an error concealment method based on the DC value.

The motion vector condition determination unit 909 receives the determination result from each of the motion vector magnitude determination unit 906 and the motion vector change amount determination unit 908. Here, when the motion vector condition determination unit 909 obtains a determination result indicating that the error concealment method using the previous frame image should be selected because the motion vector change amount determination unit 908 is within a predetermined range, The determination result of the vector size determination unit 906 is prioritized and the error concealment method using the previous frame image is set as the determination result. On the other hand, when the motion vector condition determination unit 909 obtains a determination result from the motion vector change amount determination unit 908 that the error concealment method based on the DC value should be selected because it is outside the predetermined range, the motion vector magnitude determination unit a determination result 906 to the determination result.

  That is, the motion vector condition determination unit 909 issues an instruction to select an error concealment method based on the previous frame image for a block with a small motion vector change amount, and the motion vector size is large even if the motion vector change amount is large. Is instructed to select the error concealment method based on the previous frame image, and when the change amount of the motion vector is large and the size of the motion vector is large, the instruction to select the error concealment method based on the DC value is instructed. put out.

  Note that, unlike the second embodiment, the DC / motion vector condition determination unit 907 in the present embodiment example receives the determination result of the motion vector condition determination unit 909 and the determination result of the DC determination unit 903, and The determination operation is obtained by replacing the determination result of the motion vector magnitude determination unit 706 with the determination result of the motion vector condition determination unit 909 in the determination operation of the DC / motion vector condition determination unit 707 in the second embodiment. Become.

(D-2) Moving Image Data Transmission Operation and Error Concealing Operation Next, moving image data transmission operation and error concealing operation based on the moving image transmission system according to the present embodiment will be described. However, the configuration of the encoding apparatus 300 on the moving image transmission side is the same as that of the first and second exemplary embodiments, and the operation contents thereof are also the same, so the description thereof is omitted here. Hereinafter, an operation of the decoding apparatus 900 that performs a decoding process of encoded data received via a transmission path or a propagation path, particularly an error concealment operation will be described.

  When the encoded data received from the transmission path or the propagation path is input to the decoding unit 901, DC information (pixel data based on DC values), AC information (pixel data based on AC values), The motion vector information is decoded and output to the image construction unit 902. In addition, the decoding unit 901 transmits the motion vector information of each block to the motion vector interpolation unit 905. The motion vector interpolation unit 905 holds the motion vector information of each block notified from the decoding unit 901.

  When an error is detected during decoding and detected in an intra block, the decoding unit 901 outputs the data decoded so far as it is, and outputs the intra block that cannot be decoded to the motion vector interpolation unit 905 and Notify the DC determination unit 903.

  When the motion vector interpolation unit 905 receives the error detection notification in the intra block, the motion vector decoded for the corresponding intra block (when the motion vector can be decoded by the decoding unit 901) or the average value of the motion vectors of adjacent blocks Interpolated motion vectors generated from the above (when the motion vector could not be decoded by the decoding unit 901) are held as motion vector information for the corresponding block.

  Note that the motion vector interpolation unit 905 notifies the image forming unit 902, the motion vector magnitude determination unit 706, and the motion vector change amount determination unit 908 of the motion vector or the interpolated motion vector, and holds them.

  Here, it is assumed that the decoding unit 901 detects an error during decoding and detects it in the intra block. In this case, the decoding unit 901 outputs the data that has been decoded so far as it is, and notifies each of the motion vector interpolation unit 705 and the DC determination unit 703 of an intra block that cannot be decoded. This notification is also notified to the motion vector change amount determination unit 908 via the motion vector interpolation unit 905.

  Upon receiving such notification, the motion vector change amount determination unit 908 determines whether or not the motion vector change amount for the block falls within a predetermined range. The motion vector condition determination unit 909 is notified to select an error concealment method based on the previous frame image.

  On the other hand, if it is determined that the change amount of the motion vector for the block is not within a predetermined range, the motion vector change amount determination unit 908 moves so as to select an error concealment method based on the DC value. The vector condition determination unit 909 is notified.

  On the other hand, as in the second embodiment, the motion vector magnitude determination unit 906 determines an error concealment method based on the magnitude of the motion vector and notifies the motion vector condition determination unit 909.

  The motion vector condition determination unit 909 gives priority to the determination when the motion vector change amount determination unit 908 notifies the user to select an error concealment method based on the image of the previous frame, and selects the error concealment method based on the DC value. When the notification is received, the DC / motion vector condition determining unit 907 is notified of the notification from the motion vector magnitude determining unit 906.

  The DC / motion vector condition determination unit 907 operates in the same manner as in the second embodiment to determine the final error concealment method. That is, the DC / motion vector condition determination unit 907 receives the notification of the error concealment method from each of the DC determination unit 903 and the motion vector condition determination unit 709, and selects the error concealment method based on the previous frame image from the motion vector condition determination unit 709. When the determination result that the determination should be made is given priority over the determination result of the DC determination unit 903, and the determination result that the error concealment method based on the DC value should be selected from the motion vector condition determination unit 909 is obtained The DC determination unit 903 operates to prioritize the determination result.

  Thus, the DC / motion vector condition determination unit 907 selects an appropriate error concealment method considering the DC value of the block portion where the error is detected, the magnitude of the motion vector, and the amount of change of the motion vector.

(D-3) Effects of Third Embodiment As described above, if the decoding apparatus 900 proposed in this embodiment is used, the following effects can be obtained in addition to the same effects as those of the second embodiment. it can. That is, in the present embodiment, even if the motion vector is large, it is possible to reliably select the error concealment by the previous frame image for which high correlation is expected for a block with a small change amount between the previous frame itself. Therefore, it is possible to make it difficult to output an image whose subjectivity is impaired.

(E) Fourth Embodiment In the fourth embodiment, a moving image transmission system including the encoding device 300 shown in FIG. 3 and the decoding device 1000 shown in FIG. 10 will be described. That is, a moving image transmission system configured by a combination of an existing encoding device 300 and a new decoding device 1000 will be described.

  The decoding apparatus 1000 used in this embodiment is characterized in that the determination criteria of each determination element (DC value, motion vector magnitude, and motion vector change amount) can be adjusted.

(E-1) Configuration of Encoding Device 300 and Decoding Device 1000 The configuration of the encoding device 300 is the same as that of each of the above-described embodiments. Therefore, only the configuration of the decoding apparatus 1000 will be described below.

  The decoding apparatus 1000 includes a decoding unit 1001, an image configuration unit 1002, a DC determination unit 1003, a reference memory 1004, a motion vector interpolation unit 1005, a motion vector magnitude determination unit 1006, and a DC / motion vector condition. The determination unit 1007 includes a motion vector change amount determination unit 1008, a motion vector condition determination unit 1009, and a determination adjustment unit 1010.

Among these, the difference from the third embodiment is only the determination adjustment unit 1010. That is, the other parts are the same as those in the third embodiment.

  The determination adjustment unit 1010 is a predetermined reference used as a determination criterion in the three determination units used in the third embodiment, that is, the DC determination unit 1003, the motion vector magnitude determination unit 1006, and the motion vector change amount determination unit 1008. This is to adjust the range individually. That is, the determination adjustment unit 1010 receives an instruction from the user as a control signal, and narrows or widens the predetermined range according to the instruction content of the control signal.

(E-2) Moving Image Data Transmission Operation and Error Concealment Operation The moving image data transmission operation and error concealment operation in the present embodiment are such that the boundary of the error concealment method selected by each determination unit is finely controlled by the user. Except for adjustment, the third embodiment is the same as the third embodiment.

That is, the decryption apparatus 1000 in this embodiment also has
(A) For a block that is determined to have a small amount of motion vector change, select an error concealment method based on the previous frame image regardless of the magnitude of the motion vector of the block,
(A) For a block that is determined to have a large amount of change in motion vector and that is determined to have a small motion vector, select an error concealment method based on the previous frame image;
(C) A block that is determined to have a large amount of change in motion vector and that has a large motion vector, and a DC value decoded from the block is a peripheral (including adjacent) block. For a block determined to be greatly deviated from the average value of the DC values, the error concealment method by the previous frame image is selected,
(D) A block that is determined to have a large amount of change in motion vector and that has a large motion vector, and a DC value decoded from the block is a peripheral (including adjacent) block. For blocks determined to be close to the average value of the DC values, an error concealment method based on the DC values can be selected.

  Thus, as in the third embodiment, a natural image quality with little discomfort can be obtained.

In addition, in the case of the present embodiment, in addition to such basic performance, there is an error concealment method using the previous frame image according to the source (movie, sports, news) and attribute of the moving image to be decoded. Since the boundary condition between the selected block and the block for which the error concealment method based on the DC value is selected can be adjusted, further improvement in image quality can be expected.

  In the above description, the determination adjustment unit 1010 has been described as individually adjusting the determination criterion of each determination unit in accordance with the instruction of the control signal given to each determination unit. The determination criteria may be adjusted. For this adjustment, for example, in the determination adjustment unit 1010, table means including a set of adjustment amounts of each determination unit corresponding to the control amount of the control signal is provided, and an optimal adjustment amount for the control amount of the given control signal is provided. This can be realized by giving each set to each determination unit. Further, a relational expression that defines the relationship between the control amount of the control signal and the adjustment amount of each determination unit is stored in the determination adjustment unit 1010, and the adjustment amount of each determination unit corresponding to the control amount of the control signal is calculated by calculation. It can also be realized by a calculation method.

(E-3) Effects of the Fourth Embodiment As described above, using the decoding apparatus 1000 proposed in this embodiment, in addition to the same effects as those of the third embodiment, it is based on user preferences. The adjustment of the error concealment method selection mechanism can be realized.

(F) Fifth Embodiment In the fifth embodiment, unlike the above-described embodiment, a moving image composed of the encoding device 1100 shown in FIG. 11 and the decoding device 1200 shown in FIG. An image transmission system will be described.

  In this embodiment, information on whether encoded data to be encoded and transmitted is forcibly encoded into an intra frame or an intra block is used for selection of an error concealment method on the decoding device side. It is characterized by a point.

(F-1) Configuration of Encoding Device 1100 and Decoding Device 1200 The encoding device 1100 includes an encoding unit 1101, a reference memory 1102, an intra / inter determination unit 1103, a forced intra determination unit 1104, and a decoding Unit 1105 and a forced intra notification unit 1106. That is, the encoding device 1100 has a configuration in which the forced intra notification unit 1106 is added to the configuration of the encoding device 300 according to the first embodiment.

  The compulsory intra notification unit 1106 is used to notify the decoding apparatus 1200 that a block that is currently intra-encoded and transmitted is a block that is compulsorily intra-encoded (that is, a compulsory intra block). Is done.

  For this reason, the forced intra determination unit 1104 in this embodiment is provided with a function for notifying not only the intra / inter determination unit 1103 but also the above-mentioned forced intra notification unit 1106 that intra coding should be performed. Yes.

  On the other hand, the decoding device 1200 includes a decoding unit 1201, an image configuration unit 1202, a forced intra signal reception unit 1203, and a reference memory 1204. This decoding device 1200 is provided with a forced intra signal receiving unit 1203 in place of the DC determination unit 103 of the decoding device 100 in the first embodiment, and the error concealment method executed by the image construction unit 1202 is compulsory. The configuration is the same as that of the decoding device 100 except that it is instructed by the intra signal receiving unit 1203.

  Here, the compulsory intra signal reception unit 1203 is used to receive a compulsory intra signal notified from the encoding apparatus 1100 side. When the compulsory intra signal receiving unit 1203 receives an error detection notification from the decoding unit 1201, if the compulsory intra signal is received, the compulsory intra signal receiving unit 1203 notifies the image composing unit 1202 to select an error concealment method based on the previous frame image, If the compulsory intra signal is not received, the image composing unit 402 is notified that the error concealment method using the DC component is selected. Note that the compulsory intra signal reception unit 1203 does not instruct the image configuration unit 1202 about the concealment method regardless of the presence or absence of the compulsory intra signal when the error detection notification is not received from the decoding unit 1201.

(F-2) Moving Image Data Transmission Operation and Error Concealment Operation Also in the present embodiment example, the moving image encoding and decoding operations themselves are the same as those in the first embodiment example. The difference is that when intra coding is forcibly selected at the time of encoding a moving image, the information is separately notified to the decoding device 1200 side by the forced intra notification unit 1106.

  However, the forced intra signal notified from the forced intra notification unit 1106 is information used for concealing errors. Therefore, unless an error is detected by the decoding unit 1201, the compulsory intra signal reception unit 1203 does not participate in the decoding operation at all, and the image configuration unit 1202 has an image composed of the decoded DC component and AC component. Will be output.

  On the other hand, when an error is detected during decoding of the intra block in the decoding unit 1201 and the error detection notification is notified to the forced intra receiving unit 1203, the forced intra receiving unit 1203 receives the forced intra signal. Whether or not there is an error concealment method is determined.

  For example, when a forced intra signal is received, the forced intra signal reception unit 1203 notifies the image configuration unit 1202 that an error concealment method based on the previous frame image is selected, and when the forced intra signal is not received, The image composition unit 1202 is notified that an error concealment method based on a DC value is selected.

  Thus, the image construction unit 1202 can selectively switch the error concealment method depending on whether or not the intra block is a forced intra block. Note that since the forced intra block is generally selected regardless of the amount of generated code, it is considered that there is often a high correlation with the previous frame image, so the selection of the error concealment method of the present embodiment example It is considered that higher image quality can be obtained than when an error concealment method using DC components is selected. On the other hand, if it is not a forced intra block, it is considered that the correlation with the previous frame image is low, so the selection of the error concealment method of this embodiment is higher than the case of selecting the error concealment method based on the previous frame image. It is thought that image quality can be obtained.

(F-3) Effect of Fifth Embodiment As described above, if a moving image transmission system is constructed using the encoding device 1100 and the decoding device 1200 proposed in this embodiment, an intra code is forcibly used. For the converted block, error concealment is executed by the previous frame image, and for the intra-coded block from the viewpoint of the amount of generated code, error concealment is executed by the DC component. Therefore, it is possible to provide an image with superior subjectivity as compared with the conventional apparatus.

(G) Other embodiment examples
(1) In each of the first to fifth embodiments described above, the decoding device including a functional unit that performs error concealment processing when an error is detected in an intra block has been described. In addition to the function unit, it is preferable that a function unit for concealing errors detected in the inter blocks is also provided.

(2) In the above-described embodiment, the case where the predicted DC component value is calculated from the DC component of the surrounding (including adjacent) block has been described, but the surrounding block here has some correlation with the corresponding block. As long as the block is recognized, the candidate is included regardless of the positional relationship.

(3) In the above-described second embodiment, the case where the concealment method for the error part is determined using the determination result in the DC determination unit 703 and the determination result in the motion vector magnitude determination unit 706 has been described. The error part concealment method may be determined using only the determination result of the motion vector magnitude determination unit 706.

  In this case, an error is concealed by a previous frame image in a block having a small motion vector, and an error is concealed by an image composed of decoded DC values in a block having a large motion vector. Even in this case, the image quality can be improved as compared with the conventional apparatus that performs error concealment by either one of the methods fixedly.

  In this case, the determination result of the motion vector magnitude determination unit 706 may be directly given to the image configuration unit 202, and the DC / motion vector determination unit 707 shown in FIG. 7 is not necessary.

(4) In the third embodiment described above, the error is determined using the determination result in the DC determination unit 903, the determination result in the motion vector magnitude determination unit 906, and the determination result in the motion vector change amount determination unit 908. The case of determining the concealment method of the part has been described. (A) The concealment method of the error part is determined using the determination result in the motion vector magnitude determination unit 906 and the determination result in the motion vector change amount determination unit 908. A method, (a) a method for determining a concealment method for an error part using a determination result in the DC determination unit 903 and a determination result in the motion vector change amount determination unit 908, and (c) a determination in the motion vector change amount determination unit 908. It is also possible to apply a method for determining a method for concealing an error part using only the result.

  In the case of (a) and (b), when the error concealment by the previous frame image is selected by the motion vector change amount determination unit 908, the error due to the DC value is determined by the method in the motion vector change amount determination unit 908. If concealment is selected, the error concealment method may be switched according to the determination result of the motion vector magnitude determination unit 906 or the determination result of the DC determination unit 903.

(5) In the above-described fourth embodiment, it has been described that the control signal given to the determination adjustment unit 1010 is given by the user who uses the decoding device. The present invention can also be applied to a case where it is given from a computer. With this configuration, it is possible to control an error concealment method suitable for an image from the encoding device side.

(6) In the above-described fourth embodiment, a decoding apparatus including three determination units, that is, a DC determination unit 903, a motion vector magnitude determination unit 906, and a motion vector change amount determination unit 908, and each determination Although the case where the judgment adjustment unit 1010 that adjusts the range of the judgment criterion of the unit is combined is described, the configuration of the decoding device combined with the determination adjustment unit 1010 includes (a) only a DC determination unit, A) only having a motion vector magnitude determination unit; (c) having only a motion vector change amount determination unit; (d) a combination of a DC determination unit and a motion vector magnitude determination unit; Or a motion vector change amount determination unit, or (f) a combination of a motion vector magnitude determination unit and a motion vector change amount determination unit.

(7) In the above fifth embodiment, the case where the error concealment method is switched only by the forced intra signal receiving unit 1203 has been described. However, the decoding apparatus according to each of the first to fourth embodiments and the It is also possible to combine with modification examples.
That is, when the forced intra signal receiving unit 1203 and (a) the DC determination unit are combined, (b) when the motion vector magnitude determination unit is combined, (c) the motion vector change amount determination unit is combined. In the case of (d) combining the DC determination unit and the motion vector magnitude determination unit, (e) combining the DC determination unit and the motion vector change amount determination unit, (f) the motion vector magnitude determination unit When combining with the motion vector change amount determination unit, it can also be applied to combining (i) the configurations of (a) to (f) with the determination adjustment unit.

(8) Each of the embodiments described above is applicable to any encoding device and decoding device that use an encoding method for encoding moving image data in a block format as shown in FIG. Can do.

It is a block diagram which shows the decoding apparatus which concerns on the example of 1st Embodiment. It is a figure which shows the example of a data format. It is a block diagram which shows the encoding apparatus which concerns on the example of 1st-4th embodiment. It is a block diagram which shows the example of a conventional structure of a decoding apparatus. FIG. 5 is a block diagram illustrating an image configuration unit of the decoding device illustrated in FIG. 4. It is a block diagram which shows the image structure part of the encoding apparatus which concerns on the example of 1st, 5th embodiment. It is a block diagram which shows the decoding apparatus which concerns on the example of 2nd Embodiment. It is a block diagram which shows the image structure part of the encoding apparatus which concerns on the example of 2nd, 3rd, 4th embodiment. It is a block diagram which shows the decoding apparatus which concerns on the example of 3rd Embodiment. It is a block diagram which shows the decoding apparatus which concerns on the example of 4th Embodiment. It is a block diagram which shows the encoding apparatus which concerns on the example of 5th Embodiment. It is a block diagram which shows the decoding apparatus which concerns on the example of 5th Embodiment.

Explanation of symbols

  100, 400, 700, 900, 1000, 1200 ... Decoding device, 101, 401, 701, 901, 1001, 1201 ... Decoding unit, 102, 402, 702, 902, 1002, 1201 ... Image configuration unit, 103, 703, 903, 1003 ... DC determination unit, 300, 1100 ... encoding device, 104, 404, 704, 904, 1004, 1204 ... reference memory, 301, 1101 ... encoding unit, 302, 1102 ... reference memory, 303, 1103: Intra / inter determination unit, 304, 1104: Forced intra determination unit, 305, 1105 ... Decoding unit, 403 ... Motion vector interpolation unit, 501 ... Changeover switch, 502 ... Frame data construction unit, 503 ... Pre-image processing unit 601 801... Pre-image processing unit 602 802 AC image construction unit 03, 803 ... DC image construction unit, 604, 804 ... Addition unit, 605, 805 ... Error concealment method control unit, 606, 806 ... Changeover switch, 607, 807 ... Frame data construction unit, 705, 905, 1005 ... Motion vector Interpolation unit, 706, 906, 1006 ... motion vector magnitude determination unit, 707, 907, 1007 ... DC / motion vector condition determination unit, 908, 1008 ... motion vector change amount determination unit, 909, 1009 ... motion vector condition determination unit DESCRIPTION OF SYMBOLS 1010 ... Determination adjustment part, 1106 ... Compulsory intra notification part, 1203 ... Compulsory intra signal receiving part.

Claims (2)

In a decoding device comprising decoding means for decoding and outputting encoded data that has been predictively encoded in block units,
When an error is detected for an intra-coded block during the decoding operation of the decoding means, a signal for notifying that the encoded data is forcibly intra-coded from the encoding device Check whether the signal is input, and if the signal is input, determine the error concealment process referring to the previous image. If the signal is not input, based on the decoded DC value Forced intra signal receiving means for deciding error concealment processing;
If the decoding means is performing a decoding operation normally, an original image is constructed based on the decoded data, and an error is detected for an intra-coded block during the decoding operation of the decoding means. In this case, a decoding apparatus comprising: an image composing unit that constructs an image of a corresponding part based on a concealment processing method notified from the compulsory intra signal receiving unit. In order to suppress error propagation, a compulsory intra determination unit that determines the execution of a compulsory intra encoding process as appropriate at the time of encoding, and a compulsory intra encoding process decision are notified by the compulsory intra determination unit An encoding device comprising a compulsory intra notification means for transmitting a compulsory intra signal for notifying that the corresponding encoded data is encoded data that has been compulsorily intra-encoded to the decoding device;
A transmission system comprising: the decoding device according to claim 1 .

Images