JPH0275247A - Variable rate picture hierarchy encoding system - Google Patents

Abstract

PURPOSE:To eliminate the need for an accumulator for power and information quantity for each picture element by applying hierarchical processing of an MSP(Most significant parts) and an LSP(Least significant parts) according to number of coefficients decided separately depending on the quantity of an in-block power and an inblock code information quantity in the case of applying orthogonal conversion hierarchy encoding to a moving picture signal. CONSTITUTION:An information being the result of orthogonal conversion encoding to a difference signal between adjacent frames by an orthogonal conversion coder 203 is subject to hierarchical processing into parts giving much effect on the picture quality, that is, MSP and parts giving less effect on the picture quality] that is, LSP at a hierarchy section, an identifier of disabled abort is added to the MSP and an identifier of enabled abort is added to the LSP and only the LSP with the enabled abort identifier added thereto is aborted at the network congestion. The number of orthogonal conversion coefficients assigned to the MSP in the orthogonal conversion block is increased/decreased in response to the code information quantity or the power in block. Thus, no accumulator in the unit of picture elements is required.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical field to which the invention pertains) The present invention provides a variable rate image encoding/encoding method using a high-speed packet network, which prevents deterioration of image quality even if images are discarded into bucket 1 due to network congestion. Reduce variable sea 1 to image floor M
') It concerns the encoding method.

(Prior art) - Generally, in high-efficiency encoding methods, the difference signal between adjacent frames of consecutive image frames is encoded, so if the movement of the captured image is large, a large amount of information is generated. If the movement is small, the amount of information generated is small.

In order to transmit this encoded information via a communication network with a constant transmission speed, a buffer memory is provided, and conventionally, the transmission speed is constant and the amount of encoded information stored in the buffer memory is constant. Therefore, it is necessary to control the encoding parameters, which has the disadvantage that the image quality fluctuates.

FIG. 3 is a block diagram of the transmitting side in a variable rate image encoding system intended for a communication network where the transmission speed can be made variable, such as a high-speed packet network, in order to solve such drawbacks.

This means that the input moving image signal 100 is subtracted from the signal of the previous frame stored in the frame memo January 02 by the subtracter 101, and only these difference signals are subjected to orthogonal transform encoding and vector quantum processing in the encoding unit 103. Redundancy is suppressed by encoding methods such as .

The information encoded by the encoding unit 10:3 is decoded by the decoding unit 104, added to the signal of the previous frame read from the frame memory 102 by the adder 105, and the second signal is added to the frame memory 11. Write to 02.

In this way, the frame memory 102 is updated to the latest information.

The information encoded by the encoder 103 is sent to the packet I.
It is also sent to the sending unit 106, and is sent to the high-speed bucket 1 to network 1 in units of one image frame, or in units of one block or several blocks when one image frame is divided into blocks.
It is converted into a format suitable for 07 and sent. .

In this case, if the high-speed packet I-network can ideally transmit encoded information with a short delay regardless of the amount of information generated, the image quality will be constant. However, when Pakeso 1 to discard occurs due to the net width ring, the difference between adjacent frames 11 (
Since fj and j+ are converted into the same symbol, images 11L after being discarded will not be accurately generated.

To avoid this, it is possible to retransmit the discarded packets, but this has the drawbacks of increasing delay time, lacking real time plus 1, and further increasing network speed.

Furthermore, as a layering means in orthogonal degree layering to solve these drawbacks, there is also a layering means that keeps the S/N or code information amount constant in the part that has a large effect on image quality. requires an accumulator and a buffer for each pixel, which has disadvantages such as increased hardware scale and hierarchization processing time.

(Object of the Invention) The object of the present invention is to provide a variable sea 1-image layered encoding method that suppresses the influence of discarding to bucket 1 on image quality due to network width congestion, which is a problem when transmitting moving image signals over high-speed packet networks. Our goal is to provide the following.

(Structure of the Invention) (Characteristics of the Invention and Differences from the Prior Art) The present invention provides information obtained by orthogonally transform-encoding the difference signal between adjacent frames to a portion (hereinafter referred to as M
S P (Most significant par
ts))) and the part with small influence (hereinafter referred to as L S P (
lcast;sjgnjf, 1cant parts)
), add discardable identifiers to M S P, r and ps, and discard only d and sp to which discardable identifiers are added when network @ branching. In this case, the number of orthogonal transform coefficients assigned to the MSP within the orthogonal transform block is increased or decreased according to the power within the block or the amount of code information. Its main feature is to obtain image quality.In contrast, conventional orthogonal technology requires an accumulator and buffer for each pixel to keep the S/N or code information amount constant in areas that have a large effect on image quality. This differs from the transform layered encoding method in that these are not required.

(Embodiment) FIGS. 1(a) and 1(b) are block diagrams of a transmitter and a receiver, respectively, of orthogonal transform layered coding according to an embodiment of the present invention, in which the transmitter (a) are a subtracter 201 and a frame memory 202 . Orthogonal transform encoder 203° layering unit 204. Orthogonal transform decoder 205. It is made up of an adder 206, a packet processor 1, and a sending unit 207. Further, the receiving unit (b) includes buckets 1 to 208 . Orthogonal transform decoder 209.2] 0° frame memory 211. The adder consists of adder 2]2.2]3.

Next, to explain the operation, the input moving image signal 100 is subtracted from the signal of the previous frame stored in the frame 11 memory 202 in the subtracter 201 on the transmitting side (a).
The difference signal is sent to orthogonal transform encoder 203.

The image signal encoded by the orthogonal transform encoder 20; The image signal decoded in the orthogonal transform decoder 205 and the output of the frame memory 202 are added in an adder 206, and the signal is added to the frame as the latest information. 11 memory 202.

On the other hand, MSp,r layered in the layering section 204
, sp are sent to the packet sending unit 207.

In this packet sending unit 207, bucket 1 to frame unit, 6th place in block, or line block 91
In addition, the MSP has a non-discardable identifier, and the LS
A discard iiTm identifier is added to P, and the high-speed packet network 1
07.

On the other hand, on the receiving side (b), buckets 1 to 1 received from the high-speed packet network 107 are sent to buckets 1 to 208, and
MSP, LSP (7) Identification is performed, and this MSP, LSP
are sent to the respective orthogonal transform decoders 209 and 210. Then, the MS decoded by the orthogonal transform decoder 209
I) The signal is added to the signal stored in the frame memory 211 in an adder 212, and the signal is stored in the frame memory 211 as the latest information.

On the other hand, the signal is sent to the adder 213, and the orthogonal transform decoder 210 adds r and SP that have arrived without being discarded to the decoded signal, and the moving image signal 100 is restored.

Figure 2 shows the scanning method for one block of coding coefficients after orthogonal transform coding.
Although 8 pixels are used, the present invention is also applicable to other division methods such as 16×16 pixels.

In the layering unit 204 on the transmitting side (a) in FIG.
1? It is known that when hierarchizing into MSPs and SPs, it is best to allocate MSPs from the left corner of an 8x8 pixel block.
An example is shown in which scanning is performed in the order indicated by the arrows and zero coefficients (1≦n≦64) are assigned to MSP, but the order of scanning does not necessarily follow the order of FIG. 2(a).

In layering, first, the intra-block power p of each block
Calculate. Next, two thresholds pt are applied to the intra-block power p.
ht, pthz (pth+, (pth2) are provided,
When the value of the intra-block power p is (p≦pth]), the coding coefficients of 03 (intermediate line part in Fig. 2 (b)) are set to M
Let SP and the rest be 1 and SP. When the intra-block power p is (pt≦pth2), there are 02 (Fig. 2 (
c) The coefficient of the middle plowing line part)) is set as MSP, and the rest are F and SP.
shall be. When the intra-block power p is (pth2<p), the n1 (interpolated line portion in FIG. 2(d)) encoding coefficients are set as MSP, and the rest are set as r and sp. (However, n 1 < n 2 < n :+) In this example, the intra-block power threshold value is set to two values, but a method of fixing the number of MSP coefficients without providing a threshold value, and a layer with a single threshold value are also available. It also includes hierarchization methods that provide 73 or more threshold values.

Furthermore, although this embodiment uses a method for determining the number of Msp coefficients based on intra-block power, a similar method for determining the number of Msp coefficients is also possible using intra-block code information amount. That is, a threshold is set for the amount of code information in the block, and the number of MSP coefficients is determined based on the relationship between the amount of information in the block and the threshold.

(Effects of the Invention) As described above, when the present invention performs orthogonal transform layered encoding of a moving image signal, MSP and LSP are
Since the hierarchization is performed, there is no need for an accumulator for power and information amount for each pixel, and there is an advantage that the hierarchization processing time can be shortened.

[Brief explanation of the drawing]

Fig. 1 is a block configuration diagram of an embodiment of the orthogonal transform layered coding method of the present invention, Fig. 2 is a diagram showing a method of dividing coefficients after orthogonal transform coding to explain the present invention, Fig. 3 The figure is a block diagram of a conventional variable ray 1 encoding system. 100...Moving image signal, 107 -High speed bucket 1 to network, 201...Subtractor, 202.211...Frame memory, 203...Orthogonal transform encoder, 204
...Hierarchization unit, 205.209.210 Orthogonal transform decoder, 206.2]2.2]3 ・Adder, 20
7... Bucket 1 to sending unit, 208... Packet receiving unit, n11 n21 n3 ''' Number of orthogonal transform coefficients to be Msp. Patent applicant Nippon Telegraph and Telephone Corporation Figure 2 (d) pth2<p

Claims (2)

[Claims] (1) The difference signal between the input video signal and the image signal stored in the frame memory is subjected to orthogonal transform encoding, and the encoded image signal is hierarchically divided into parts that have a large effect on image quality and parts that have a small effect on the image quality. , the frame memory is rewritten with an image signal obtained by decoding only the portion with a large influence, and is stored as the latest image signal, and a non-discardable identifier is added to the portion with a large influence, and a discardable identifier is added to a portion with a small influence. The packets are multiplexed and transmitted, and on the receiving side, the parts with a large influence and the parts with a small influence are identified, and each is orthogonally transformed and decoded. The image signal of the part with a large influence is rewritten in the frame memory and stored as the latest image signal. , variable rate image hierarchical encoding characterized in that a portion having a small effect on image quality that is not discarded in the orthogonal transform decoding is added to the decoded image signal to restore the moving image signal. method. (2) When hierarchizing the image signal into parts that have a large effect on image quality and parts that have a small effect, many orthogonal transform coefficients are 2. The variable rate image hierarchical encoding method according to claim 1, wherein a small number of orthogonal transform coefficients are assigned to blocks having small intra-block power or small intra-block code amount, since the influence on image quality is small.