JP3024380B2 - Image coding control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to control of data encoding and decoding processes used in, for example, video conferences and video telephones.

[0002]

2. Description of the Related Art Generally, image encoding processing technology is described in "Atsumi Murakami's High Efficiency Coding Technology" (Television Society Journal V
ol. 42 # 11P. 1198-),
It can be roughly classified into information source coding processing technology and transmission line coding processing technology. In the information source coding processing technology, a moving image is regarded as a continuation of still images (hereinafter, referred to as frames), and a block in which digital image data forming a frame is grouped into a plurality of adjacent units is used as a unit of the coding process. By applying various coding processes in combination, highly efficient compression of the information amount of the image is achieved. On the other hand, the transmission path coding process is performed when a plurality of coded data obtained as a result of various coding processes are actually transmitted, and each coded data is converted into a variable length code corresponding to the frequency of occurrence. Substitute and serial multiplex these.

[0003] The operation of an image transmission apparatus for encoding and decoding image data will be described below as an example of a conventional technique. FIG. 9 is a configuration diagram of a general image transmission apparatus that simplifies and expresses the contents of the above-mentioned paper, where 1 denotes A /
D conversion section, 2 format conversion section, 3 information source coding section, 4 transmission path coding section, 5 transmission control section, 6 transmission path, 7 reception control section, 8 transmission path decoding section, 9 is an information source decoding unit, 10 is a format reverse conversion unit, and 11 is a D / A conversion unit.

In FIG. 9, an analog television signal is digitized by an A / D converter 1 and further divided by a format converter 2 into blocks suitable for information source coding. In the information source coding unit 3, various coding processes such as motion compensation, discrete cosine transform, and quantization are performed on the block of digital image data, and the block is developed into various coded data. These various types of encoded data are entropy-encoded and multiplexed by the transmission path encoding unit 4 and transmitted to the transmission path 6 as an image encoded bit string under the control of the transmission control unit 5. At this time, the transmission speed of the image coded bit string corresponds to the communication speed of the transmission path 6.
The communication speed of the transmission line 6 is generally fixed. The reception control unit 7 transmits an image coded bit string from a communication partner to the transmission path 6.
, And the transmission line decoding unit 8 separates and decodes various encoded data. The information source decoding unit 9 decodes the block of digital image data in the reverse procedure of the encoding process in the information lower encoding unit 3 based on various encoded data,
Further, the format inverse converter 10 performs deblocking, and restores the analog television signal via the D / A converter 11. The above is the outline of the encoding process in the general image transmission device. In addition, as a general image encoding processing method, CCITT SGXV Report
R37'Recommendation of th
e H-seriesH. 261 ′ (hereinafter, referred to as H261).

[0005] Next, the operation of the transmission path encoding unit 4 and the transmission path decoding unit 8 will be described in detail below. Transmission path coding unit 4
In the above, when generating an image coded bit string, various kinds of coded data are replaced with variable length codes corresponding to the frequency of occurrence, and these are serially multiplexed, so that the code amount of the image coded bit string per one image frame Varies depending on the properties of the encoding target frame. In particular, when the motion compensated prediction inter-frame coding represented by the above H261 is applied, the above tendency is remarkably exhibited. In order to efficiently transmit the image coded bit string, it is essential that the image coded bit string be buffered on the transmission side and the reception side of the image transmission apparatus. However, the transmission side buffer (TB
F) and the processing delay of the entire system of the image transmission device increases in proportion to the accumulation amount of the image coding bit string in the reception side buffer (RBF). FIG. 7 is a functional block diagram around the transmission path coding section, and FIG. 8 is a functional block diagram around the transmission path decoding section. In the figure, 3 is an information source coding section, 4 is a transmission path coding section, 41, 42, 4
3 is a variable length code converter, 44 is a serial multiplexer, 45 is a TBF, 46 is a transmission path coding control unit, 5 is a transmission control unit,
7 is a reception control unit, 8 is a transmission path decoding unit, 81 is an RBF, 8
2 is a code word separator, 83, 84, and 85 are variable-length code inverse converters, 86 is a transmission path decoding control unit, and 9 is an information source decoding unit.

In FIG. 7, when the above-mentioned H261 is used as an image coding method, for example, the information source coding unit 3 performs motion compensation prediction, discrete cosine transform, scalar quantization, etc. on the image data to be coded. And outputs the scalar-quantized discrete cosine transform coefficients, motion compensation information, and quantization characteristics as encoded data. The transmission path encoding unit 4 converts each of the encoded data into a variable length code converter 41,
At 42 and 43, it is independently replaced with a code word. Here, a unique variable length code whose code length is determined according to the frequency of occurrence is used as the code word. Further, the serial multiplexer 44
Performs serial multiplexing of the variable-length code of each encoded data in block units under the control of the transmission path encoding control unit 46. The image encoded bit string obtained as described above is temporarily stored in the TBF 45 and then transmitted to the transmission path via the transmission control unit 5. On the other hand, in FIG. 8, the reception control unit 7 sends out only the image coded bit sequence in the bit sequence received from the transmission line 6 to the transmission line decoding unit 8. The transmission line decoding unit 8 temporarily stores the image coded bit sequence in the RBF 81, and then extracts the necessary image coded bit sequence from the RBF 81 by the codeword separator 82 under the control of the transmission line decoding control unit 86. Codewords corresponding to the coded data (for example, discrete cosine transform coefficients after scalar quantization, motion compensation information, and quantization characteristics) are separated and input to variable-length code inverse transformers 83, 84, and 85, respectively. The variable-length code inverse converter 8
3, 84 and 85 replace the codeword input with the corresponding coded data and output to the information source decoding unit 9.

Here, since the code amount for each image frame in the image coded bit string is not uniform, the amount of accumulation in the TBF 45 at the end of the coding of each frame varies. Furthermore, depending on the amount of code per frame of the image, the TBF 45 may overflow or underflow. This is the same for the RBF 81. When an overflow occurs in the TBF 45 and the RBF 81, a part of the image coded bit string to be input to the buffer is discarded, and the operation of the system is broken. Also, when the TBF 45 underflows, a bit sequence other than the image coded bit sequence is transmitted to the transmission path, so that the transmission efficiency of the image coded bit sequence is reduced.
Further, when the RBF 81 underflows, the decoding process is performed intermittently, so that the frame interval of the encoding target image cannot be faithfully reproduced, and frame dropping occurs.

As an example of the countermeasure for overflow / underflow of the TBF 45, there is the following processing. That is,
The transmission path coding control unit 46 monitors the accumulated amount of the TBF 45, stops the coding process for the information source coding unit 3 immediately before the overflow, and notifies the transmission control unit 5 immediately before the underflow. This is a process of notifying the image encoding bit string transmission invalidity. Information source coding unit 3 that has received the coding process stop
Immediately stops the encoding process and waits until the suspension of the encoding process is released. In addition, the transmission control unit 5 that has received the image coded bit string transmission invalid transmits “a bit string other than the image coded bit string” to the transmission path 6 that can be identified and discarded on the receiving side. However, the former causes a reduction in the number of frames of a transmission image, and the latter causes a reduction in transmission efficiency.

In the transmission line decoding unit 8, the time required to separate and decode various kinds of coded data for one image frame from the coded bit sequence fluctuates irregularly. This is due to the fact that the code amount per frame and the number of symbols generated per image frame in the image encoded bit string are not equal. This impedes the faithful reproduction of the frame interval of the image to be encoded and causes dropouts. The above countermeasures include separating and decoding various kinds of encoded data from an image encoded bit sequence, and further buffering the encoded data.
However, the buffer capacity required for the countermeasure is enormous as compared with the RBF 81 that stores the image coded bit string without redundancy, and the device scale increases.

[0010] In order to solve the above problems, some restrictions are required for buffering of the image coded bit string on the transmission side and the reception side of the image transmission apparatus. Encoded bit string 1
Upper limit of code amount per frame and RBF of standard pseudo decoder
Only the capacity is specified, and no exact constraints are given.

[0011]

In a conventional transmission apparatus, the buffering of a coded bit stream requires TB
Overflow or underflow of F and RBF occurs, which causes a reduction in the transmission efficiency of the coded bit sequence and, in the case of an image transmission device, causes deterioration in visual characteristics of a reproduced image due to dropping of frames. This has been a problem in applying image coding to “high quality and applications that do not allow dropout of frames” typified by digital transmission and recording / accumulation of broadcast quality images at a fixed rate.

The present invention prevents overflow / underflow of TBF and RBF in a small-scale device,
An object is to prevent a reduction in transmission efficiency of an encoded bit string.

[0013]

An image code according to the present invention
The transmission control method uses an image data consisting of multiple frames on the transmitting side.
Bit string consisting of multiple codewords
Is generated, and the generated encoded bit sequence is temporarily
Buffer and send it to the receiving side.
Whether the coded bit string is received and stored in the receiving buffer
Image coding control method on the transmitting side when decoding from
Where the average decoding per codeword at the receiving side is
Signal time to t, buffering delay of the sending buffer
The time is defined as the frame period T of the image data to be encoded, and 1 frame.
Per frame when the frame is divided into small areas
When the total number of small areas of A is A
The number of codewords at the time of encoding on the transmitting side such that the maximum number of codewords Smax to be obtained is Smax ≦ T / (A × t).
Is controlled.

An image coding control system according to the next invention
Indicates that the transmitting side encodes image data consisting of multiple frames.
To generate an encoded bit sequence consisting of a plurality of codewords,
Once the generated coded bit string is stored in the sending buffer,
And then send it to the receiver, where the receiver
Receive the data sequence, store it in the receiving buffer, and then decode
Image encoding control method on the transmission side at the time,
Encode the buffering delay time of the above transmission side buffer
The frame period T of the target image data, the transmission side buffer
The output speed when outputting a coded bit string from
The generated code amount of the i-th frame after activation on the transmission side is E _i ,
, The maximum storage capacity ^Be _{max of the} transmission side buffer
The, B ^e _max ≧ 2RT set, the accumulation amount of the transmission side buffer B ^e _i, or on
The generated code amount E _i for the i-th frame is
Is, frame for encoding in ^{_{RT ≦ B e i ≦ B e}} max -RT 2RT-B e i-1 ≦ E i ≦ B e max -B e i-1 and comprising as the transmission side
It is characterized in that the generated code amount for each is controlled.

Further , in these inventions, the start-up
If the receiver operates before the
The bit rate for receiving the coded bit string is R,
The buffering delay time of the buffer
Frame period T of the
Where the transmission delay time is L,
A capacity B ^e _max, the maximum capacity B ^d _max of the receiving-side buffer
The, respectively, and sets the _{^{2RT ≦ B e max ≦ R (}} L + T) B d max x ≧ R (L + T).

Further , in these inventions, the starting of the transmitting side is performed.
After receiving, the receiving side operates and the first
Less than one frame up to the beginning of the frame
Discards the incompletely coded bit string
When the receive buffer accumulates the encoded bit string from the position
If the receiving side receives the encoded bit sequence,
Is R, the buffering delay time of the transmission side buffer
Is the frame period T of the image data to be encoded,
Assuming that the buffering delay time of the buffer is L,
The maximum capacity ^Be _max of the sending buffer and the receiving buffer
The maximum capacity B ^d _max of §, respectively, and sets the _{^{2RT ≦ B e max ≦ R (}} L + T) B d max ≧ R (L-T) + B e max.

[0017]In addition, in these inventions, the starting of the transmitting side is performed.
After the operation, the receiving side operates and the receiving side buffer operates.
Stores all received image coding bit strings once.
If the receiving side receives the coded bit string,
Set the cut rate to R, buffering the sending buffer
The delay time is calculated based on the frame period T of the image data to be encoded.
The buffering delay time of the receiving buffer is (L +
T), as the maximum capacity B of the transmission side buffer
^e _max And the maximum capacity B of the receiving buffer. ^d _max And
Each, 2RT ≦ B ^e _max ≤R (L + T) B ^d _max ≧ RL + B ^e _max Is set.

[0018]

[0019]

[0020]

[0021]

[0022]

The detecting means detects information related to the encoding of the encoding means, and the control means controls the encoding of the encoding means based on the detection of the detecting means.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An image transmission apparatus as an embodiment of a data encoding / decoding processing control apparatus according to the present invention will be described below. This control method is mainly applied to a transmission path encoding unit and a transmission path decoding unit among the functional blocks of the image transmission apparatus. The operation of both functional blocks will be described below. FIG. 1 is a functional block diagram around a transmission line coding unit, and FIG. 2 is a functional block diagram around a transmission line decoding unit. In the figure, 3 is an information source coding unit, 31 is a code amount / symbol number control unit, 4 is a transmission line coding unit, 41, 42, and 43 are variable length code converters, 44 is a serial multiplexer, and 45 is a TBF. , 46 are a transmission path coding control unit, 47 is a code amount / symbol number counting unit, 5 is a transmission control unit, 51 and 71 are communication procedure control units, 7 is a reception control unit, 7
2 is a frame head position detector, 8 is a transmission path decoder, 81
Is an RBF, 82 is a codeword separator, 83, 84, and 85 are variable-length code inverse converters, 86 is a transmission line decoding control unit, 87 is a timer, and 9 is an information source decoding unit.

The operations of the transmission line coding unit 4 and the transmission line decoding unit 8 according to the present invention will be described below only with respect to differences from the operation in the description of the prior art. In FIG. 1, a code amount / symbol number counting unit 47 monitors an image coding bit string output from the serial multiplexer 44, totals the code amount and the number of symbols in real time, and outputs the totaled result as the code amount / symbol number. It returns to the control unit 31. The code amount / symbol number control unit 31 changes the quantization characteristic based on the totaling result so that the code amount / symbol number for the frame currently being encoded satisfies a predetermined constraint. Thus, the information source coding unit 3 is controlled. In particular, the number of symbols can be controlled in units of small areas smaller than the frame. At the same time, when the encoding of the same frame is completed, the constraint condition of the code amount / symbol number for the next frame is obtained. The communication procedure control unit 51 prepares a transmission start condition of the image coded bit string by transferring the control signal to the same functional block provided on the receiving side. In FIG. 2, a communication procedure control unit 71
Controls, for example, to prohibit transmission of the image encoded bit sequence until the receiving side can receive the image encoded bit sequence. The frame head position detection unit 72 monitors the image coded bit string received by the reception control unit 7 and notifies the transmission path decoding control unit 86 of the head position in the image coded bit string. By the way, the actual processing time required for separating / decoding a codeword per frame from an image coding bit string varies depending on the code amount and the number of symbols of the same frame.
In order to reproduce the frame interval in the image to be encoded, the processing time is calculated by setting the frame period of the image to be encoded to T.
Needs to be corrected. A timer 87 is used as a reference for the correction. For example, if the time correction is managed in frame units, if the separation / decoding time of the code word for one frame is less than T, the shortage time is set by the timer 8.
7 and prohibits the image coded bit string from being taken out from the RBF 81 during that period. However, it is assumed that the separation / decoding time of the code word for one frame is T or less. The various encoded data separated and decoded as described above are sequentially sent to the information source decoding unit 9 and decoded by the same functional block. The encoding means includes an information source encoding unit 3 and variable length code converters 41, 42,
43, the transmission-side storage means is composed of the TBF 45, the transmission control means is composed of the transmission control unit 5 and the communication procedure control unit, and the reception control means is the reception control unit 7, the communication procedure control unit The receiving side storage means is composed of RBF 81, the decoding means is composed of the information source decoding section 9 and the variable length code inverse transformer, the detecting means is the code amount / symbol number counting section, and the control means is the code And an amount / symbol number control unit.

As a first embodiment, a constraint on the number of symbols of coded data generated in the image data processing of the information source coding unit of the image transmission apparatus is derived, and the information source coding is performed based on this constraint. The encoding of the image data processing of the unit is controlled. It is assumed that the transmission path decoding unit 8 corrects the codeword separation / decoding processing time in small area units obtained by equally dividing a frame into “A”. Correction target time per small area: τ = T / A (T: frame period of image to be coded) The “small area divided into“ A ”” is a codeword separation / decoding time on the receiving side. A management unit for correction, A
The case of = 1 corresponds to the case where the time correction is managed on a frame basis. By the way, assuming that the average decoding time per symbol of a code word is t, the symbol number control unit performs an arithmetic process on an upper limit: Smax of the number of symbols generated per the small area. Smax = τ / t = T / (A × t)
(1) The number-of-symbols control unit inputs the total number of small words per frame and the frame period of the image to be coded, which are units of codeword separation / decoding processing time correction.
The information source coding unit 3 is controlled so that the number of symbols generated per small area is equal to or smaller than Smax.

Next, as a second embodiment, a change in the storage capacity of the TBF 45 will be considered. Here, it is assumed that the information source encoding unit 3 and the transmission line encoding unit 4 can perform encoding processing on the encoding target image without delay. Further, the TBF 45 is flushed before the start of the encoding process, and after the start of the encoding process, after securing a buffering delay of period: T, reading at a fixed bit rate: R starts. Here, T is the frame period of the image to be encoded. Also,
The TBF buffering delay is a delay in the encoding processing unit for compensating for a case where the generation of the image encoding bit sequence for the encoding target frame is concentrated near the end of the same frame period. The image encoding bit sequence for the encoding target frame occurs in a burst within the same frame period. Note that the degree to which the image coded bit string is generated differs depending on the characteristics of the image (such as the pattern and the movement of the subject) and the configuration of the coding processing unit. FIG. 3 shows a time-series change in the TBF accumulation amount shown in consideration of the above (the first frame to be encoded after the start of the encoding process is the first frame).
The hatched portions in the figure indicate regions where the amount of accumulated TBF can be taken. The constraint on the amount of code generated by the information source coding unit is given as an overflow / underflow prevention condition for the TBF 45. Here, · E _i: the code amount of the i-th frame (for convenience, and E ₀ = 0) (input from the serial multiplexer 44) · B ^e _i: TBF accumulation at time iT (for convenience, B
^e ₀ = RT to) (Input from TBF45) · B ^e _max: When TBF volume (calculated from the parameters of the fixed system), an overflow / underflow preventing conditions TBF45 than 3, RT ≦ B ^e _i ≦ B ^e _max -RT (2) Thus, whereas B ^e _max ≧ 2RT,

[0027]

(Equation 1)

[0028] ^{_{(3) E i = B e}} i -B e i-1 + RT (4) (2) from equation equation from equation _{^{(4), 2RT-B e i-1}} ≦ E i ≦ B e max - from B ^e _i-1 or more, due to overflow / underflow prevention TBF45 is, B ^e _max ≧ 2RT made to _{^{set, RT ≦ B e i ≦ B}} e max -RT (5) 2RT-B e i- ₁ ≦ E _i ≦ ^Be _max −B ^e _i−1 (6)

The operation of the receiving side when the above-described symbol number control and code amount control are performed on the transmitting side of the image transmission apparatus will be described below. Here, the information source decoding unit 9
It is assumed that the decoding process per frame of image data is executed without delay at the frame period T on the transmission side. Further, the RBF 81 is flushed when the receiving side is started up, and the reception control unit 7 accumulates a bit stream output at a fixed bit rate: R corresponding to the line speed without delay, and secures a predetermined buffering delay before transmitting. It is assumed that an image coded bit sequence for one image frame is output within a frame period T on the transmission side under the control of the path decoding control unit 86. Note that the extraction of the image encoded bit sequence from the RBF 81 is performed in a burst within the frame period: T depending on the configuration of the variable length code inverse transformer. Under the above conditions, the following three cases will be considered in which the storage capacity of the RBF 81 varies.

(First Condition) First, the first condition will be described. It is assumed that the receiving side becomes operable before the transmitting side of the image transmission apparatus is activated. This corresponds to, for example, the normal reproduction of the image storage medium, and in the image transmission apparatus, the communication procedure control unit 5
This is realized by the control of 1,71. Note that the reception control unit 7 detects and discards bit strings other than the image coding bit string. FIG. 4 shows a time-series change in the accumulated amount of RBF when a buffering delay of period: L is set in the RBF 81 (the decoding start time of the first image frame after starting the receiving side is set to 0). The hatched portions in the figure indicate regions where the amount of accumulated RBF can be taken. here,
A new constraint condition regarding buffering is derived from the overflow / underflow prevention condition of the RBF 81. B ^d1 _k : RBF accumulation amount at time kT (here, B
^d1 ₀ = RL) B ^d1 _max : RBF capacity: System fixed parameter

[0031]

(Equation 2)

[0032] (2), (7) from the equation, R (L + 2T) -B e max ≦ B d1 k ≦ RL (8) from 4, due to overflow / underflow preventing RBF81 the RT ≦ B ^d1 _k ≦ B ^d1 _max −RT (9) From the equations (8) and (9), the new constraint conditions are as follows. _{^{· 2RT ≦ B e max ≦ R}} (L + T) ( provided that L ≧ T) (10) · B d1 max ≧ R (L + T) (11)

(Second Condition) Next, the second condition will be described. When the receiving side becomes operable after the transmitting side of the image transmitting apparatus is started, an incomplete image coded bit string of less than one frame from immediately after the start of reception to the first detected frame is converted to an RBF.
It is assumed that the storage target 81 is not to be stored. This is an example of coping with a situation that is scattered in an ordinary image transmission apparatus, and aims to reduce the total buffering delay by accumulating only the image encoding bit string to be decoded in the RBF 81. When discarding the image coded bit string of the incomplete frame, the frame head position in the image coded bit string is detected by the frame head position detection unit 72,
The result is notified to the transmission path decoding control unit 86 to identify the bit string to be discarded. In addition, the reception control unit 7 discards bit strings other than the image coding bit string. Now, the RBF when a buffering delay of period: L is set in the RBF 81
FIG. 5 shows a time-series change in the accumulation amount (the decoding start time of the first image frame after starting on the receiving side is 0, and the serial number of the same frame on the transmitting side is c + 1). The hatched portions in the figure indicate regions where the amount of accumulated RBF can be taken. Here, the overflow of RBF81 /
A new constraint on buffering is derived from the underflow prevention condition. B ^d2 _n : RBF accumulation amount at time n T (here, B
^d2 ₀ = RL) B ^d2 _max : RBF capacity: a parameter fixed to the system

[0034]

(Equation 3)

[0035] (2), (12) from equation, R (L + T) + B e c -B e max ≦ B d2 n ≦ R (L-T) + B e c ie, R (L + 2T) -B e max ≦ B ^d2 _n ≦ R (L−2T) + B ^e _max (13) From FIG. 5, to prevent the overflow / underflow of the RBF 81, RT ≦ B ^d2 _n ≦ B ^d2 _max −RT (14) Equation (13) From equation (14), the new constraints are as follows. _{^{· 2RT ≦ B e max ≦ R}} (L + T) ( provided that L ≧ T) (15) · B d2 max ≧ R (L-T) + B e max (16) The above constraints, the image code It is also effective when the receiving side of the general-purpose application device becomes operable before the transmitting side is activated.

(Third Condition) Further, the third condition will be described. It is assumed that the receiving side becomes operable after the transmitting side of the image transmitting apparatus is activated, and that all the received image encoded bit strings are to be stored in the RBF 81. As described above, this is an example of coping with a situation that is scattered in a normal image transmission apparatus.
The control of the RBF 81 is simplified. The image coded bit string of the incomplete frame is immediately discarded when the codeword separator 82 accesses the RBF 81. In addition, the reception control unit 7 discards bit strings other than the image coding bit string. Now, the RB when the buffering delay of period: X is set in the RBF 81
FIG. 6 shows a time-series change of the F accumulation amount (the decoding start possible time of the first complete image frame after activation on the receiving side is set to 0, and the serial number of the same frame on the transmitting side is c +
1). The hatched portions in the figure indicate regions where the amount of accumulated RBF can be taken. Here, a new constraint condition regarding buffering is derived from the overflow / underflow prevention condition of the RBF 81. · E _c ^pt: code quantity · B ^d3 _m (c-th frame on the transmission side) should be discarded incomplete frame: RBF accumulation at time mT (where, B
^d3 ₀ = RX) B ^d3 _max : RBF capacity: system fixed parameter

[0037]

(Equation 4)

[0038] Here, equation (2) and _{(4), 0 <(B c -E c} pt) <B e max -RT (18) (2) and Equation (17), (18) more _{R (X + T) -E c} pt + B e c -B e max ≦ B d3 m ≦ R (X-T) + B e c -E c pt i.e., R (L + X) + B e max <B d3 m <R ( X−2T) + B ^e _max (19) From FIG. 6, in order to prevent the overflow / underflow of the RBF 81, RT ≦ B ^d3 _m ≦ B ^d3 _max −RT (20) From the equations (19) and (20), _{^{· 2RT ≦ B e max ≦ RX}} ( However, X ≧ a ^{2T) (21) · B d3} max ≧ R (X-T) + B e max (22) By the way, as the X = L + T, and rearranging the results,
The new constraints are: _{^{· 2RT ≦ B e max ≦ R}} (L + T) ( provided that L ≧ T) (23) · B d3 max ≧ RL + B e max (24) · RBF buffering delay: (L + T) The above constraints This is also effective when the receiving side of the image coding application device becomes operable before activation of the transmitting side. Similarly, the above constraint also applies when the receiving side of the image coding application device becomes operable after activation of the transmitting side and the image coding bit sequence of the incomplete frame is excluded from the RBF accumulation target.

As described above, according to the present invention, the decoding / decoding time of the code word becomes uniform for each of the A small areas obtained by equally dividing the frame by "A" on the image coding application equipment decoding side. On the transmitting side, the number of symbols generated per A small areas: Smax is restricted by the following equation (1) so that the separation / decoding time of the codeword for one frame is fixed. Since it is guaranteed that the value becomes T, it is possible to realize image encoded data transmission without dropping frames with a small-capacity buffer. Smax ≦ T / (A × t) (1) (where T: frame period of image data to be encoded t: average decoding time per symbol of codeword)

On the transmitting side of the image coding application device, the code amount per one frame of the image is calculated by the following equation (5),
By giving the constraint of (6), overflow / underflow does not occur in the TBF, and it is possible to prevent system operation failure and transmission efficiency reduction. _{^{B e max ≧ 2RT RT ≦ B}} e i ≦ B e max -RT (5) 2RT-B e i-1 ≦ E i ≦ B e max -B e i-1 (6) ( however, T: coded frame period of the image data R: constant bit rate corresponding to the line speed E _i: i-th generated code amount for the frame B ^e _i: TBF accumulation amount B ^e _max at the end the frame coding: TBF volume)

When the receiving side of the image coding application device becomes operable before the start of the transmitting side, the buffering delays of the TBF and the RBF are set to T and L, respectively.
The following equations (10) and (1) are respectively applied to the F capacity and the RBF capacity.
By imposing the constraint of 1), it is possible to realize image coded data transmission at a constant frame rate without causing a system operation failure or a reduction in transmission efficiency. _{^{2RT ≦ B e max ≦ R (}} L + T) (10) B d1 max ≧ R (L + T) (11) ( _{^{although, L ≧ T B e max:}} TBF capacity B ^d1 _max: RBF volume)

When the receiving side of the image coding application device becomes operable after activation of the transmitting side, and the image coding bit string of the incomplete frame is excluded from the RBF accumulation target, TB
Assuming that the buffering delays of F and RBF are T and L, respectively, the following restrictions are applied to each of the TBF capacity and the RBF capacity, so that the image coded data transmission at a constant frame rate can be performed without causing system operation failure or reduction in transmission efficiency. It can be realized without getting up. Note that the following equations (15) and (1)
The constraint condition 6) is also valid when the receiving side of the image coding application device becomes operable before activation of the transmitting side. _{^{2RT ≦ B e max ≦ R (}} L + T) (15) B d2 max ≧ R (L-T) + B e max (16) ( _{^{although, L ≧ T B e max:}} TBF capacity B ^d2 _max: RBF volume)

Further, when the receiving side of the image coding application device becomes operable after the start of the transmitting side, and all the received image coding bit strings are to be stored in the RBF, the TBF
By setting the buffering delays of TBF and RBF to T and L + T, respectively, and imposing the following restrictions on the TBF capacity and the RBF capacity, image coded data transmission at a constant frame rate causes a system operation failure and a reduction in transmission efficiency. It can be realized without. Note that the following constraints are also valid when the receiving side of the image coding application device becomes operable before activation of the transmitting side. Similarly, the following equations (23) and (24) also apply to the case where the image encoding application device receiving side becomes operable after activation of the transmitting side and the image encoding bit sequence of the incomplete frame is excluded from the RBF accumulation target.
The above conditions apply. _{^{2RT ≦ B e max ≦ R (}} L + T) (23) B d3 max ≧ RL + B e max (24) ( _{^{although, L ≧ T B e max:}} TBF capacity B ^d3 _max: RBF volume)

[0044]

As described above, the present invention detects information relating to the data encoded by the encoding means, and controls the encoding processing of the encoding means based on the detection result, thereby achieving high-efficiency transmission. Becomes possible.

[Brief description of the drawings]

FIG. 1 is a functional block diagram around a transmission path coding unit in an image transmission apparatus using the present invention.

FIG. 2 is a functional block diagram around a transmission path decoding unit in the image transmission device using the present invention.

FIG. 3 is a diagram showing a time-series change of a TBF storage capacity according to the present invention.

FIG. 4 is a diagram showing a time-series change of the RBF storage capacity according to the present invention when the receiving side of the image coding application device becomes operable before activation of the transmitting side.

FIG. 5 is a diagram illustrating a time-series change of the RBF storage capacity according to the present invention, in which the image coding applied device receiving side becomes operable after the transmitting side is activated, and the image coding bit string of the incomplete frame is excluded from the RBF storage target. FIG.

FIG. 6 is a diagram showing an example in which the receiving side of the image coding application device becomes operable after the start of the transmitting side, and the received image coding bit sequence is changed to R
FIG. 9 is a diagram illustrating a case where a BF is to be stored.

FIG. 7 is a functional block diagram around a transmission path coding unit in a conventional image transmission apparatus.

FIG. 8 is a functional block diagram around a transmission path decoding unit in a conventional image transmission device.

FIG. 9 is a functional block diagram of a general image transmission device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 3 information source coding unit 4 transmission line coding unit 5 transmission control unit 7 reception control unit 8 transmission line decoding unit 9 information source decoding unit 31 code amount / symbol number control unit 41 variable length code converter 42 variable length code converter 43 Variable Length Code Converter 44 Serial Multiplexer 45 TBF 46 Transmission Line Coding Control Unit 47 Code Amount / Symbol Counting Unit 51 Communication Procedure Control Unit 71 Communication Procedure Control Unit 72 Frame Head Position Detector 81 RBF 82 Codeword Separator 83 variable-length code reverse converter 84 variable-length code reverse converter 85 variable-length code reverse converter 86 transmission line decoding control unit 87 timer The same reference numerals in the figure indicate the same or corresponding parts.

Continuation of front page (72) Inventor Shinichi Okada 5-1-1, Ofuna, Kamakura City Mitsubishi Electric Corporation Communication Systems Laboratory (56) References JP-A-3-148981 (JP, A) (58) Survey Field (Int.Cl. ⁷ , DB name) H04N 7/24 G06T 9/00 H03M 7/30

Claims (5)

(57) [Claims] An image data comprising a plurality of frames is transmitted on a transmitting side.
Bit string consisting of multiple codewords
Is generated, and the generated encoded bit sequence is temporarily
Buffer and send it to the receiving side.
Whether the coded bit string is received and stored in the receiving buffer
Image coding control method on the transmitting side when decoding from
And the average decoding time per codeword on the receiving side is
t, the buffering delay time of the transmitting buffer
Frame period T of image data to be encoded
Small area per frame when divided into areas
Where A is the total number of
The number of codewords at the time of encoding on the transmitting side so that the large codeword number Smax satisfies Smax ≦ T / (A × t).
An image coding control method characterized by controlling the following. 2. An image data comprising a plurality of frames on a transmitting side.
Bit string consisting of multiple codewords
Is generated, and the generated encoded bit sequence is temporarily
Buffer and send it to the receiving side.
Whether the coded bit string is received and stored in the receiving buffer
Image coding control method on the transmitting side when decoding from
A is, when the buffering delay of the transmission side buffer
Frame over arm period T of the encoding target image data between the coded bit stream from the transmission side buffer
The output speed at the time of output is R, i-th after the start of the transmitting side
When the generated code amount of the frame of E is assumed to be E _i ,
The maximum storage capacity B ^e _max side buffer is set to B ^e _max ≧ 2RT, accumulation amount B ^e _i of the transmission side buffer, and the i-th
Generating for the frame code amount E _i, respectively, in ^{_{RT ≦ B e i ≦ B e}} max -RT 2RT-B e i-1 ≦ E i ≦ B e max -B e i-1 and comprising as the transmission side Frame for encoding
Image coding characterized that you control the generated code amount for each
control method. 3. The receiving side operates before the transmitting side starts.
If the receiving side receives the coded bit sequence
Rate is R, buffering delay of the above sending buffer
The time is defined as the frame period T of the image data to be encoded.
Assuming that the buffering delay time of the receiving buffer is L,
The maximum capacity ^Be _max of the transmitting buffer and the receiving buffer
3. The maximum capacity B ^d _max of the buffer is set to 2RT ≦ B ^e _max ≦ R (L + T) B ^d _max x ≧ R (L + T) , respectively.
Image coding control method described above. 4. The receiving side operates after the transmitting side is activated,
Immediately after the start of reception
Of incompletely coded bit strings less than one frame
Discarded, and the encoded bit string from the beginning of the frame is
If the receive buffer accumulates, the receiving side
R is the bit rate for receiving the data stream,
The buffering delay time of the
Frame period T, buffering delay of the receiving buffer
Assuming that the delay time is L, the maximum capacity B of the transmission side buffer
and ^e _max, the maximum capacity B ^d _max of the reception buffer, its
^{_{Respectively, 2RT ≦ B e max ≦ R}} (L + T) B d max ≧ R (L-T) + B e claim and sets the _max 1 or claim 2 Symbol
Image coding control method described above. 5. The receiving side operates after the transmitting side is activated.
All image encodings received by the receiving buffer
If the bit string is stored once, the receiver
R is the bit rate for receiving the encoded bit sequence,
The buffering delay time of the buffer is
Frame period T of the data, buffer of the receiving side buffer
Assuming that the ring delay time is (L + T),
A maximum capacity B ^e _max fa, the maximum capacity of the reception buffer
And the amount B ^d _{max, ^{respectively, 2RT ≦ B e max ≦ R}} (L + T) B d max ≧ RL + B e claim and sets the _max 1 or claim 2 Symbol
Image coding control method described above.