JPH0298292A - Variable rate picture gradation coding system - Google Patents

Abstract

PURPOSE:To improve picture quality by providing a threshold level to an in- block power when a moving picture signal subjected to orthogonal conversion is subjected to gradation, comparing the in-block power with the threshold level and varying (in-block code quantity assigned to MSP in-block total code quantity) adaptively. CONSTITUTION:In the case of gradation, at first an in-block power measuring device 403 and an in-block total code quantity measuring circuit 404 calculate in-block power (p) of each block and in-block total code quantity (c). Then the in-block power (p) is compared with two in-block power threshold values pth1, pth2 (pth1<pth2) provided in advance. Then when the in-block power (p) is expressed as p<=pth1, the number of orthogonal conversion coefficients assigned to the MSP to obtain the n1% of the in-block total code quantity (c) to be in the in-block code quantity of the part MPS giving a large effect and the remaining coefficients are assigned to the LSP giving a small effect. Thus, the picture quality is improved.

Description

[Detailed Description of the Invention] "Industrial Application Field" This invention reduces deterioration in image quality even if packets are discarded due to congestion in a variable rate image encoding method that uses high-speed bucket IQ. This relates to a layered encoding method.

``Prior Art'' FIG. 4 is a diagram showing the transmitting side of a variable rate image encoding system intended for a communication network such as a high-speed packet network that can vary the transmission speed, and the input moving image signal is , subtractor 1
01 is subtracted from the signal of the previous frame stored in frame memo January 02, and only these difference signals are sent to the encoder 1.
In step 03, redundancy is suppressed by an encoding method such as orthogonal transform encoding or vector quantization. The information encoded by the encoding unit 103 is decoded by the decoding unit 104, and the information is decoded by the adder 10.
5, the signal is added to the signal of the previous frame read from the frame memory 102, and this signal is written into the frame memory 102. In this way, the frame memory 102 is updated to the latest information. The information encoded by the encoding section 103 is also sent to the packet sending section 106, where it is converted into a format suitable for a high-speed packet network in units of several blocks and sent out.

In this case, if the high-speed packet network ideally can transmit encoded information with a short delay regardless of the amount of information generated, the image quality will be constant. However, when a packet is discarded due to network comparison, the image after the discard is not accurately reproduced because the difference signal between adjacent frames is encoded. In order to avoid this, it is possible to retransmit the discarded packets, but this has the drawbacks of increasing delay time, lack of real-time performance, and further increasing network congestion.

In order to solve this drawback, the image frame is divided into blocks, and within that block, the parts that have a large effect on image quality (hereinafter referred to as MSP: Mo5t51gn1-
ficaat Parts) and the least affected part (hereinafter referred to as LSP: Least 51gn1ficant Pa
rts), collect several blocks worth of MSPs and LSPs, and configure packets consisting only of MSPs and packets consisting only of LSPs. During transmission, packets consisting of MSPs are given a non-discardable identifier, and packets consisting of LSPs are assigned a non-discardable identifier. A variable rate image coding method (Japanese Patent Application No. 1652-1983) in which a discardable identifier is added to the LSP, and when network congestion occurs, packets consisting of LSPs to which the discardable identifier is attached are discarded.
15) has been proposed. In an example of orthogonal transform layered encoding to explain this method, when layering MSP and LSP, orthogonal transform coefficients are scanned in a zigzag manner to obtain the amount of encoded information, and a preset value (for example, block Coefficients up to 75% of the total code amount per block are assigned to MSP, and other coefficients are assigned to LSP. However, if a packet containing LSP of a block with high movement is discarded due to network congestion, decoding In some cases, a situation may arise in which it is not possible to restore an undegraded image with the amount of code for MSP, while there may be considerable redundancy because an excessive amount of code is allocated to MSP for blocks with small motion. There is a problem.

SUMMARY OF THE INVENTION An object of the present invention is to provide a variable rate image layered encoding method that suppresses the influence on image quality of packet discards due to network congestion, which is a problem when transmitting moving image signals over a high-speed packet network.

``Means for Solving the Problems'' This invention provides information obtained by orthogonally transform-encoding the difference signal between adjacent frames to a portion (MSP) that has a large impact on image quality.
During transmission, several blocks worth of MSP and LSP are collected, and MSP
A non-discardable identifier is added to a packet consisting of MSP, a discardable identifier is added to a packet consisting of LSP, and a discardable identifier is added to a packet consisting of LSP during network comparison. In orthogonal transform layered coding transmission that discards only packets consisting of added LSPs, during layering, the intra-block power is measured for each orthogonal transform block, and the intra-block power is set to a predetermined intra-block power threshold. For blocks with intra-block power greater than or equal to the threshold, the intra-block code amount ratio allocated to MSP is increased, and for blocks with intra-block power less than the threshold, the intra-block code amount ratio allocated to MSP is increased. This technology is mainly characterized by adaptive orthogonal transform layered coding that reduces
This differs greatly from the conventional orthogonal transform layered coding method, in which there is a high possibility that the amount of transmitted codes for blocks with large intra-block power will be insufficient, or that the amount of transmitted codes for blocks with small intra-block power will be excessive.

Embodiment FIGS. 1(a) and Φ) are block diagrams of a transmitter and a receiver of an orthogonal transform layered coding transmission system to which the present invention is applied, respectively, including a subtracter 201, a frame memory 202,
Orthogonal transform encoder 203, layering section 204, orthogonal transform decoder 205, adder 206, packet sending section 207, packet receiving section 208, orthogonal transform decoders 209 and 210, frame memory 211, and adders 212 and 213. Become.

The input image signal is subtracted from the previous frame signal stored in the frame memory 202 in the subtracter 201, and a difference signal is sent to the orthogonal transform encoder 203. The image signal encoded by the orthogonal transform encoder 203 is processed by the layering unit 204 to divide the image signal into a portion (M
SP) and a portion having a small influence on image quality (LSP), and only the MSP is sent to the orthogonal transform decoder 205, and the image signal decoded by the orthogonal transform decoder 205 and the output of the frame memory 202 are The signals are added in an adder 206 and stored in the frame memory 202 as the latest information.

On the other hand, the MSP is hierarchized by the hierarchization unit 204.

The LSP is sent to the packet sending unit 207, which adds a non-discardable identifier to the packet consisting of MSP and a discardable identifier to the packet consisting of LSP, and sends the packet to the high-speed packet network.

On the receiving side, the packet received from high-speed packet m is sent to MSP in packet receiving section 208.

The LSP is identified, and the MSP and LSP are sent to orthogonal transform decoders 209 and 210, respectively. Orthogonal transform decoder 2
The signal decoded by 09 is stored in the frame memory 211
The adder 212 adds this signal to the signal stored in the frame memory 211 as the latest information. On the other hand, the signal is sent to the adder 213, and the LSP that has reached the orthogonal transform decoder 210 without being discarded is
is added to the decoded signal to restore the moving image signal.

Hereinafter, the present invention will be described in detail with reference to FIGS. 2 and 3, taking as an example the case where there are two intra-block power thresholds for hierarchization.

FIG. 3(a) shows a scanning method for one block of coding coefficients after orthogonal transform coding, and in the figure, l block is 8 x 8 pixels, but the present invention has 16 x 16 pixels. It is also applicable to other block division methods. In the layering unit 204 in FIG. 1, when layering into MSP and LSP, 8×8 orthogonal transform coefficients are
It is known that it is good to allocate SP, and in this embodiment, an example is shown in which scanning is performed in the order indicated by the arrows in FIG. 3(a).
The order does not depend on the order shown in Figure (a).

FIG. 2 is a block diagram showing the hierarchization unit 204 in FIG. 1 in detail, and shows the intra-block power threshold input device 40.
1,,? ISP code amount ratio manual device 402, intra-block power measurement circuit 403, total intra-block code amount measurement circuit 4
04 and a hierarchical circuit 405. Prior to transmission,
Intra-block power threshold input device 401. Using the MSP code amount ratio input device 402, the intra-block power threshold pthlSpth2 for layering and the MSP code amount ratios nl (%), n2 (%), and n3 (%) are input and set.

In layering, first, the intra-block power measuring device 403
, the intra-block total code amount measuring circuit 404 calculates the intra-block power p and the intra-block total code amount C of each block. Next, the intra-block power p and two preset intra-block power thresholds pthl and pth2(pt
hl < pth2), and when the value of the intra-block power p is (p≦pthl), n of the total intra-block code amount C
The number of orthogonal transform coefficients to be assigned to MSP is determined so that 1% becomes the intra-block code amount of MPS, and the remaining coefficients are
Assign to SP. (Figure 3(b)).

Similarly, the intra-block power p is (pthl<P≦pt
h2) In the case of (pth2 < p), n2% and n3% of the total intra-block code amount C become the intra-block code amount of MSP, as shown in FIGS. 3(C) and (b), respectively. The number of orthogonal transform coefficients to be assigned to MSP is determined as follows, and the remaining coefficients are assigned to LSP. (However, n
1<n2<n3) Also, for example, the distribution of the intra-block power p is measured in advance, and the threshold values pthl and pLh2 are set as follows.
The probability of (p≦pthl) and the probability of (pLh2<p) are determined to be approximately equal, and nl,,n
2. If n3 is defined as n2=(nl+n3)/2, the amount of MSP transmission codes will be approximately equal to the amount of transmission codes when the MSP code amount is uniformly transmitted at n2%, but LSP discards Regarding the image quality when this occurs, compared to the case where the MSP code amount is uniformly transmitted at n2%, the image quality is improved by the amount that the image quality of the block with a large intra-block power p is improved. (For blocks with small intra-block power p, redundancy is simply lost and there is no deterioration in image quality.) In this example, two intra-block power threshold values are used, but one threshold value is used. It is clear that similar techniques can also be used to implement the hierarchization method described above and the hierarchization method in which three or more threshold values are provided.

"Effects of the Invention" As explained above, when orthogonally transformed moving image signals are hierarchized, a threshold is set for the intra-block power, and the intra-block power p is compared with the intra-block power threshold. (M
By adaptively changing the intra-block code amount/total intra-block code amount allocated to SP, when an LSP packet is discarded, the transmission code amount is increased in blocks with large intra-block power, improving image quality. On the other hand, in blocks with small intra-block power, redundancy can be achieved, so by setting the intra-block power threshold and the intra-block code amount ratio to appropriate values, it is possible to allocate to conventional MSP. Compared to the case where the ratio of the intra-block code amount is made uniform, this method has the advantage that image quality can be improved with the same amount of code.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the orthogonal transform layered encoding method, FIG. 2 is a block diagram showing a specific example of the layering section 204 in FIG. 1, and FIG. 3 is a block diagram for explaining this invention. FIG. 4 is a block diagram showing a conventional variable rate encoding method. 301 Orthogonal transform coefficient region containing n1% of the total code amount within the second block 302 Orthogonal transform coefficient region containing n2% of the total code amount within the second block 303 Orthogonal transform coefficient region containing 03% of the total code amount within the second block 1 Figure patent applicant Nippon Telegraph and Telephone Corporation

Claims (1)

[Claims] (1) Orthogonally transform the inter-claim difference signal of the video signal,
Encodes in block units, and layers the encoded information into parts that have a large impact on image quality and parts that have a small impact, and for each part, combines several blocks such as line blocks into one packet, Effects on image quality in a variable rate image layered encoding method that adds a non-discardable identifier to packets that consist only of parts, and a discardable identifier to packets that consist of only parts that have a small effect on image quality, and multiplexes and transmits the packets. In order to hierarchize into large and small parts, a threshold is set in advance for the power within the block that has undergone orthogonal transformation, and during hierarchization, the power within the block is measured for each block, and the power within the block is set as the threshold. Due to the size relationship with the value, for blocks with large intra-block power, the allocated code amount ratio to the total intra-block code amount allocated to the part that has a large effect on image quality is increased, and for blocks with small intra-block power, the allocated code amount ratio is increased. A variable rate image hierarchical encoding method characterized by layering that reduces the ratio of allocated code amount to the total intra-block code amount allocated to a large portion.