JPH03184475A - Transform encoding system - Google Patents

Abstract

PURPOSE:To widely reduce processing time and the amount of information to be generated by providing a means to count the number of continuous zero coefficients out of quantized transform coefficient values, means to determine a threshold value for the number of the continuous zero coefficients to be transmitted based on the amount of data caused to remain in a transmission buffer, and means to make this quantized outputs zero to the following transform coefficients when the count value exceeds the threshold value. CONSTITUTION:An encoding control part 7 determines an quantizing step size 109 from a buffer remaining amount 108 of a transmission buffer 6. At a threshold value setting part 9, a threshold value 111 is determined similarly from the buffer remaining amount 108 and outputted to a decision part 10. A counter 8 counts how many zero coefficient values are continuous, and when the coefficient value not zero is outputted from a quantizing part 11, a count value 110 is reset and becomes zero. At the decision part 10, the count value 110 is compared with the threshold value 111 and a compared result 112 is outputted to the quantizing part 11. When the decided result 112 shows the count value = the threshold value, quantization is not executed but the coefficients from the next coefficient to the final coefficient are outputted as zero coefficients.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to band compression of image data using a linear transform coding method [Prior art] Fig. 3 shows, for example, W, H, CHE N, W,
K, PRATT, ”Scene
eA d a p t i v e Cod e r
”, (IEEE Transactions onc.
communications, vol, cOM-3
2, No. 2, No. 3, 1984). In the figure, (1)
(2) is a blocking unit that blocks the input signal; (2) is a linear transformation unit that linearly transforms the blocked signal; (3)
is a scan converter that rearranges the signal sequence within the block, (
4) is a quantization section, (5) is an encoding section, (6) is a transmission buffer, and (7) is an encoding control section. Next, the operation will be explained. For one digitized input image signal (101) for one frame, the blocking unit (1) converts n pixels in the horizontal and vertical directions (n is a natural number, for example, n
=4.8.16) into two-dimensional blocks. A linear transformation unit (2) performs two-dimensional linear transformation (for example, orthogonal transformation such as discrete cosine transformation) on the blocked image signal (102) to generate a transform coefficient block (103) in the spatial frequency domain. Here, for example, 8×8 pixel block f(x, y)(x+y=0+L”
', 7) is given by the following equation. Here, u, v=o, 1. ..., 7, where x and y are coordinates in the pixel domain, and u and v are coordinates in the transformation domain. Transform coefficient block F (u * v) (u + v
The properties of = 0+ 1*...+ 7) will be explained based on FIG. The value of F(u,v) indicates the level of each spatial frequency component included in the blocked image signal (102). The frequency in the horizontal direction increases as the value of U increases, and the frequency in the vertical direction increases as the value of ■ increases. In other words, the value of F(0,0) is the block image signal (102
) corresponds to the intensity of the DC component, and the value of F(7, 7) corresponds to the intensity of the AC component having high frequencies in both the horizontal and vertical directions. Therefore, for a flat image block such as a background with little change in pixel values, only low frequency components have non-zero significant coefficients, and high frequency components have almost zero coefficients. Conversely, for image blocks such as edge portions where pixel values change rapidly, non-zero significant coefficients are assigned to high frequency components as well as low frequency components. Next, in the scan conversion unit (3), the conversion coefficient block (1
03), the transform coefficients are rearranged in the order indicated by the arrows in FIG. 4, for example, and a transform coefficient sequence (104) is output. In the case of the previous 8×8 pixel block, a coefficient string of 64 coefficients for one block is output. The sorting is done by scanning in a zigzag pattern from the conversion coefficients of low frequency components where significant coefficients are likely to occur to the conversion coefficients of high frequency components where significant coefficients are difficult to occur, so that the significant coefficients are placed in the first half and the string of zero coefficients is continued as long as possible in the second half. I do it to make it work. Next, the quantization unit (4) quantizes the transform coefficient sequence (104) with a given quantization step size (109) and outputs a quantized coefficient sequence Q(u,v) (105). Encoding section (
In step 5), a code is assigned to the quantized coefficient sequence (105) and output as encoded data (106) to the transmission buffer (6). Here, two-dimensional variable length rod coding will be described as an example of a code assignment method. This combines the number of consecutive zero coefficients (hereinafter referred to as a run) for the quantized coefficient sequence (105) and the quantization level of the non-zero coefficients that follow, and then calculates 1 for the combined event (run, level). This is done by assigning two Huffman codes. For example, in the case of the quantized coefficient sequence shown in FIG. 5, the events (runs, levels) are as follows. (0,20), (2,15), (4,5), (3,2)
, (7, 1), EOB Here, EOB is a mark indicating that there are no non-zero coefficients thereafter and zero coefficients continue until the end of the block. Therefore, in the case of this quantized coefficient sequence, a predetermined Huffman code is assigned to each of the six events including EOB. Next, the transmission buffer (6) smoothes the fluctuating amount of generated information and sends it out to the transmission path (107) at a constant rate. The encoding control unit (7) adaptively controls the quantization step size (109) from the buffer remaining amount (108), which is the remaining data amount in the transmission buffer (6), and outputs it to the quantization unit (4). do. In other words, the remaining buffer capacity (108
), the input image is coarsely quantized by increasing the quantization step size (109) in order to reduce the amount of information generated. Conversely, when the remaining buffer capacity (108) is small, the input image is finely quantized by reducing the quantization step size (109) in order to increase the amount of information generated. Note that since the number of nonzero coefficients varies greatly depending on the size of the quantization step size (109), the amount of information generated varies greatly from 11110 times to several thousand times.

[Problem M that the invention attempts to solve]

Conventional transform coefficient encoding methods are careful as described above, so not only do all coefficients have to be quantized, but the amount of information generated fluctuates greatly depending on the size of the quantization step size. The problem was that it was difficult to control the rate of change. This invention was made to solve the above-mentioned problems, and it determines the transmission range of coefficients according to the quantized coefficient sequence in the transform block, shortens the processing time required for quantization, and improves information generation. The purpose of this invention is to obtain a transform coefficient encoding method that reduces the amount of variation. [Means for Solving Problem In the quantized conversion method, means for counting the number of consecutive zero coefficients among the values of the quantized transform coefficients, and means for determining a threshold value for the number of consecutive zero coefficients to be transmitted based on the remaining amount of data in the transmission buffer. , and means for setting the quantized output of the subsequent transform coefficient to zero when the counted value of zero coefficients exceeds a determined threshold value. [Operation] The transform coding method according to the present invention linearly transforms the input signal block, quantizes it in a given order, counts the number of consecutive zero coefficients, and determines the number from the remaining amount of data in the transmission buffer. When the threshold value is exceeded, the value of the subsequent quantization coefficient is set to zero. Embodiment 1 of the Invention An embodiment of the invention will be described below with reference to FIG. In the figure, (8) is a counter that counts consecutive zero quantization coefficients, (9) is a threshold setting unit that sets a threshold value, and (10
) is a determination unit that compares the count value and a threshold value and makes a determination; (11
) is a quantization unit whose operation is changed by a determination circuit, and the other parts are the same as in FIG. 3. Next, the operation will be explained. Similar to FIG. 3, the digitized input image signal for one frame (101) is subjected to linear transformation and rearrangement of the transformation coefficients), then to the quantization unit (11).
Quantization is performed using the quantization step size (109) given by , and a quantization coefficient sequence (105) is output. The encoding control unit (7) controls the remaining buffer capacity (108
), the quantization step size (109) is determined and output to the quantization unit (11). Further, the threshold value setting section (9) similarly determines a threshold value (111) from the remaining buffer capacity (10B) and outputs it to the determination section (10). The counter (8) counts how many consecutive zero coefficient values are present in the quantized coefficient sequence (105), and the non-zero coefficient values are counted by the quantization unit (11).
), the count value (110) is reset to zero. The determination unit (10) calculates the input count value (11
0) and the threshold (111) given from the threshold setting unit (9)
The determination result (112) is sent to the quantization unit (11
). In the quantization unit (11), the determination result (112
), the process branches to the following. (1) When the determination result (112) is equal to the count value or the threshold value, the next coefficient is also quantized with the same quantization step size (109). (II) When the determination result (112) is a count value of two thresholds, no quantization is performed and the coefficients from the next coefficient to the last coefficient are output as zero coefficients. FIG. 2 shows the flow of these operations. Further, in the example of FIG. 5, for example, when the threshold value is set to 4 or 5, the events and the number of coefficients to be quantized are as follows. When the threshold is 4, the events are (0, 20), (2, 15), EO
B, and the number of coefficients to be quantized is eight, from Q(0,0) to Q(2,1). When the threshold value is 5, the events are (0, 20), (2, 15), (4
,5), (3,2), EOB, and the number of coefficients to be quantized is 18 from Q(0,0) to Q(3,2). As mentioned earlier, the intensity of the transform coefficient generally becomes weaker as it goes from low frequency to high frequency component, so the probability that the quantized coefficient Q (u, v) of the quantization result becomes zero continuously becomes higher as the frequency becomes higher. expensive. Therefore, if we quantize in order from low frequency to high frequency components and count the zero quantization coefficients, when the threshold (111) is small, the condition (II) above is satisfied for the coefficients of relatively low frequency components. , the quantization and encoding of the coefficient conversion coefficients of high-frequency precepts can be aborted, and the amount of calculations and amount of information generated can be reduced. Therefore, in the past, the amount of information generated was controlled only by the quantization step size (109), and the amount of information generated changed rapidly depending on the size of the quantization step size (109). By controlling this, it is possible to suppress rapid changes in the amount of information generated. In other words, when the remaining buffer capacity (108) is small, the threshold value (111
) by reducing the coefficient transmission range, the amount of information generated at a small quantization step size (109) can be suppressed. Conversely, the remaining buffer capacity (1
08) is large, by setting the threshold value (111) large, it is possible to transmit even the coefficients of high frequency components and prevent deterioration in the quality of the decoded image. Furthermore, according to this embodiment, it is not necessary to quantize all the transform coefficients, so when implemented using a digital signal processor or the like, the amount of arithmetic processing can be significantly reduced and further effects can be obtained. Note that in the above embodiments, a combination of two-dimensional linear transformation and quantization has been described, but the combination of two-dimensional linear transformation and quantization is only one-dimensional. A similar effect can be obtained with a combination of three-dimensional or other linear transformation and quantization. [Effect of the Invention J As described above, according to the present invention, it is determined whether or not to quantize and encode subsequent transform coefficients based on the number of consecutive zero quantization coefficients, which improves the image quality of decoded images. This has the effect of significantly reducing the processing time required for quantization and the amount of information generated without degrading the data.

[Brief explanation of drawings]

FIG. 1 is a block diagram illustrating one embodiment of the present invention, and FIG.
3 is a block diagram of a conventional example, FIG. 4 is a diagram for explaining the properties of transform coefficient blocks, and FIG. 5 is a diagram for explaining code assignment. It is a diagram. (1) Blocking unit, (2) Linear conversion unit, (3) Scan conversion unit, (4), (11) Quantization unit, (5) Encoding unit, (6) Transmission buffer, (7) is an encoding control unit, (8> is a counter, (9) is a threshold setting unit, (10)
is a determination unit, (105) is a quantized coefficient sequence, (110) is a count value, (111) is a threshold value, and (112) is a determination result. In the drawings, the same reference numerals indicate the same or corresponding parts. Agent Masu Oiwa
male 2nd diagram Q 4th diagram conversion grandchild number block F(u,v)

Claims (1)

[Claims] In a transform encoding method that linearly transforms an input signal sequence and sequentially quantizes and encodes transform coefficients from low to high frequencies in the transform domain, among the values of the quantized transform coefficients, means for counting the number of consecutive zero coefficients; means for determining a threshold value for the number of consecutive zero coefficients to be transmitted based on the remaining amount of data in the transmission buffer; and subsequent conversion when the counted value exceeds the threshold value. 1. A transform coding method comprising means for zeroing quantized outputs of coefficients.