JPH02260882A - Picture encoder - Google Patents

Abstract

PURPOSE:To attain communication for multi-point simultaneously between a decoder for a moving picture and a decoder for a still picture by providing a control means controlling a 1st code conversion means and a 2nd code conversion means. CONSTITUTION:A code converter 12 converts a prediction error signal fed from a quantizer 7 into a code when a control signal from a control circuit 4 represents the coding of a moving picture and stops code conversion when the signal represents coding of a still picture and outputs a code representing an invalid data. Moreover, coding for refresh and changeover of inter-frame code are applied by using a signal from a refresh circuit 3 and the signal subject to code conversion is outputted while being in matching with the line speed. A code converter 13 converts a prediction error signal fed from the quantizer 7 into a code when a control signal from the control circuit 4 represents the coding of a still picture and stops code conversion when the signal represents coding of a moving picture and outputs a code representing an invalid data while matching with the line speed. Thus, multi-point communication between a decoder for a moving picture and a decoder for a still picture is attained.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention uses a moving image encoding device of a moving image transmitting station to encode moving images and still images. The present invention relates to an image encoding device for performing multipoint communication with a still image receiving station.

(Prior Art) Conventionally, in the communication of images in teleconferences, etc., a video communication station communicates video images with another video communication station.
The still image communication station communicates still images with the still image communication station. Alternatively, still images are communicated between a video communication station and a video communication station using a still image mode in a video encoding device or a video decoding device.

(Problem to be Solved by the Invention) However, in conventional image communication in teleconferences, etc., the encoding processing methods of the moving image encoding device and the still image decoding device are different, so the video communication station The communication station and the still image communication station communicate with the still image communication station in their respective closed areas, and it is impossible to communicate between the moving image layer and the still image communication station, which is extremely inconvenient.

SUMMARY OF THE INVENTION In view of the above-mentioned drawbacks, the technical problem of the present invention is to provide an image encoding device that can perform multi-point communication simultaneously between a moving image decoding device and a still image decoding device. be.

(Means for Solving the Problem) According to the present invention, in an image encoding device that uses correlation within a screen and correlation between screens, means for storing one screen of input images, and means for storing the input image and the screen. a first selection means for selecting one of the output signals from the first selection means; means for obtaining a signal with reduced redundancy using the output of the first selection means and a third prediction signal; means for quantizing the redundancy-reduced signal, which when used can set the redundancy-reduced signal to a predetermined value according to a third instruction; and a means for quantizing the redundancy-reduced signal; means for obtaining a locally decoded signal using the output; a first predictor that generates a first prediction signal from the locally decoded signal using intra-picture correlation; and a second predictor using inter-picture correlation; a second predictor capable of generating an n1 signal and storing the second predicted signal according to the second instruction; second selection means for selecting according to instructions and generating the third p prediction signal;
a first code converting means capable of converting the code of the output of the quantizing means and converting the code as invalid regardless of the input according to the second instruction; , a second code conversion means capable of performing code conversion by invalidating all prediction errors as a result of using the correlation between screens according to the second instruction; and a first control means that performs the first instruction; giving the second instruction to the quantizing means, the second predictor, the first code converting means, and the second code converting means, and giving a third instruction to the quantizing means; An image encoding device characterized by comprising two control means is obtained.

(Summary of the Invention) The invention of the present application enables communication of still images between a moving image encoding device and a still image decoding device, which was previously impossible. It performs multi-point image communication with the image communication station.

That is, as shown in FIG. 1, in order to perform multi-point communication between the moving image encoding device 40, the moving image decoding device 41, and the still image single-image decoding device 42, the moving image encoding device 40 and The moving image decoding device 41 is connected with a high speed line 43 for moving images, and the moving image encoding device 40 and the still image decoding device 42 are connected with a low speed line 44 for still images. Then, normal video communication is performed between the video encoding device 40 and the video decoding device 41, and the video encoding device 40
Still images are communicated between the still image decoding device 42 and the still image decoding device 42 . To encode a still image in the video encoding device 40, a still image to be transmitted to the video encoding device 40 is captured, the captured image is encoded little by little, and the encoded data is gradually transmitted to the transmission path. It is necessary to flow it to This is because the transmission path leading to the still image decoding device 42 has a much lower transmission rate than the transmission path leading to the moving image decoding device 41, so a large amount of data cannot be sent instantaneously. .

Therefore, it is necessary for the moving image encoding device 40 to encode still images little by little and transmit them.

Normally, as shown in Figures 2(a), (b), and (c), in video encoding, images are periodically rewritten (refreshed) to remove errors caused by errors in the transmission path. ) is being carried out. Refreshing is performed by intra-frame coding using intra-screen correlation while moving n lines at a time for each predetermined screen, removing errors that occur in the decoder's frame memory due to transmission path errors, and creating a new clean image. is written into the frame memory of both the encoder and demultiplexer. In areas other than refresh, inter-frame encoding is performed using the inter-screen boundaries.

Therefore, when still images are encoded by the video encoding device of the present invention, the refresh mode is used, and each n line is encoded by intra-frame encoding, and the areas other than refresh are encoded by inter-frame encoding so that the data is invalid. It is encoded by replacing it with a code that indicates a certain fact, thereby suppressing the generation of excessive information on each screen.

By encoding and transmitting data bit by bit in this way, a large amount of information is not generated instantly.

When encoding a still image, as shown in FIG. 3, a frame memory 43 for storing still images is prepared in front of the video encoding device 44, and the still image to be transmitted is stored in the frame memory 43. Write.

The output of the still image frame memory 43 is selected as the input to the moving image encoding device 44, and encodes n lines for each screen determined in the refresh mode.

The video encoding device of the present invention has a configuration as shown in FIG. 4, and has two predictors, an interframe predictor 45 and an intraframe predictor 46, for normal encoding and refreshing. The predictors 45 and 46 are switched. In normal encoding, the output of the interframe predictor 45 with high encoding efficiency is selected as a prediction signal, and the difference from the previous screen is encoded.

When refreshing, the period (n lines)
The output of the intraframe predictor 46 is selected as the predicted signal, and intraframe encoding is performed.

At this time, the intra-frame predictor 46 uses the value of the previous pixel or the neighboring pixel in the screen of the input signal, so it is not affected by the previous screen in which errors due to transmission path errors have been accumulated, and therefore the intra-frame predictor 46 produces a correct image. The contents of frame memory can be updated. When encoding a still image, all encoded data in areas other than the refresh line within the screen, that is, areas encoded using interframe prediction, are replaced with invalid data between frames. send to the side That is, the only valid data on the screen is the refresh line, and it is possible to suppress the instantaneous generation of a large amount of information. The encoded data at this time is converted into an efficient code such as a Huffman code by the second code converter 48, matched with the speed of the transmission line, and data is transmitted little by little over a low-speed line for still images. send. At this time, the video encoding device stores the interframe prediction signal immediately before the start of still image encoding in the interframe predictor n1 45, and while executing the still image encoding, By replacing the output of the first code converter 47 with invalid data, the moving image decoding device can freeze the image immediately before the still image is encoded.

Alternatively, a signal supplied from a still image frame memory may be used as an input signal to the moving image encoding device, and encoding may be performed as if a normal moving image signal were manually generated.

Then, when the video encoding device restarts video encoding, the prediction signal stored in the interframe predictor 45 is used to newly encode the entire screen using intra-screen correlation. Immediately after the encoding of a still image is completed, efficient video encoding can be performed using the correlation between frames without refreshing with intra-frame encoding. Similar to the video encoding device, the video decoding device also does not refresh the entire screen.
Interframe decoding can be performed immediately.

The still image decoding device on the receiving side, as shown in FIG.
By providing a feedback loop, signals sent from the video encoding device can be correctly decoded. The decoded signal selection switch selects the output of the intraframe decoder 49 when intraframe encoding is performed, outputs the data as an output of the still image decoding device, and stores it in the frame memory. The output of the interframe decoder 50 is selected when the area has been subjected to interframe encoding. In this way, for example, time t in FIG.
In the area refreshed at 11, interframe encoding is executed at time 11, and the refresh line moves to the next area, but in the area where interframe encoding was executed, the encoded data is invalid at the time of encoding. Therefore, even if interframe decoding is performed, the refreshed data stored in the frame memory becomes the output of the interframe decoder 50, and the refreshed data stored in the frame memory is retained. dripping By repeating this operation, the area to be refreshed moves, and the area of the new refreshed image is gradually enlarged to become one still image.

If the still image decoding device has a feedback loop as described above, still images can be transmitted using the refresh mode of the moving image encoding device.

(Example) Next, one embodiment of the present invention will be described with reference to the drawings.

First Embodiment FIG. 6 is a block diagram showing the configuration of an encoding apparatus according to a first embodiment of the present invention. The input image signal is line 100
The signal is supplied to the frame memory 1, sync separator 2 and switch 5 via. The frame memory 1 writes an input image signal when a write signal supplied from the control circuit 4 via a line 41 instructs writing, and stores the screen. The sync separator 2 separates sync from the input moving image signal supplied via the line 100, and generates a frame pulse indicating the beginning of a frame and a line pulse indicating the beginning of a line based on the separated sync signal. , refre・
It is supplied to the static circuit 3 and the control circuit 4. The refresh circuit 3 generates a refresh control signal based on the frame pulse and line pulse supplied from the sync separator 2, which performs refresh while moving by n lines every predetermined frame. The refresh control signal generated by the refresh circuit 3 is supplied to the control circuit 4, the switch 11, the code converter 12, and the code converter 13. $IJ! 11 circuit 4 generates a write signal for one frame time when the still picture signal supplied via line 400 instructs encoding of a still picture;
1 to the frame memory 1. Further, when the still image signal supplied via the line 400 instructs encoding of the still image, the control circuit 4 generates a control signal for executing the encoding of the still image, and connects the line 450. The signal is supplied to a switch 5, a frame memory 10, a code converter 12 and a code converter 13 via the converter. Furthermore, the control circuit 4
When the still picture signal supplied via the line 400 instructs still picture encoding and the refresh control signal supplied from the refresh circuit 3 indicates a line other than refresh, A stop signal is supplied to the quantizer 7 to stop quantization and make the data invalid.

When the control signal supplied from the control circuit 4 via the line 450 indicates encoding of a moving image, the switch 5 selects the input moving image signal supplied via the line 100 and outputs the control signal. When indicates the encoding of a still image,
Select frame memory 1 output. The output signal of the switch 5 is supplied to a subtracter 6. The subtracter 6 subtracts the signal supplied from the switch 5 and the prediction signal supplied from the switch 11 to obtain a child-side error signal.

The output signal of the subtracter 6 is supplied to a quantizer 7.

When the stop signal supplied from the control circuit 4 indicates stop, the quantizer 7 stops quantization and outputs invalid data. When the stop signal indicates that the stop signal is not a stop, the prediction error signal supplied from the subtracter 6 is quantized.

In other words, when a still image is being encoded, normal quantization is performed in areas where intra-frame encoding is performed, and data is invalidated between frames in areas where inter-frame encoding is performed. The amount of information generated by the encoding is only the refresh line that has been intra-frame encoded. When still images are being encoded in this way, generation of excessive information is suppressed by gradually encoding them little by little. The output signal of the quantizer 7 is supplied to an adder 8, a code converter 12, and a code converter 13. The adder 8 adds the quantized prediction error signal supplied from the quantizer 7 and the prediction signal supplied from the switch 1] to obtain a locally decoded signal. The locally decoded signal obtained by adder 8 is supplied to intraframe predictor 9 and frame memory 10. Frame inner aFJ device 9
generates an intra-frame prediction signal using the value of a pixel in the vicinity of the pixel to be encoded, for example, a previous pixel or a previous line, among the locally decoded signals supplied from the adder 8; Supplied to switch 11.

When the control signal supplied from the control circuit 4 indicates encoding of a moving image, the frame memory 10 delays the local decoded signal supplied from the adder 8 by one frame time, and generates an interframe predetermined signal n1. However, when the control signal indicates encoding of a still image, writing of the local decoded signal supplied from the adder 8 is stopped and the signal of the previous frame is held. The interframe prediction signal output from frame memory 10 is supplied to switch 11 . The switch 11 is
When the refresh control signal supplied from the Ritzch circuit 3 indicates a refresh line, the output of the intra-frame predictor 9 is selected; when the refresh control signal indicates a line other than the refresh line, the output of the intra-frame predictor 9 is selected. , selects the output of the frame memory 10. The predicted signal output from switch 11 is supplied to subtracter 6 and adder 8 .

The code converter 12 converts the quantizer 7 when the control signal supplied from the control circuit 4 indicates encoding of a moving image.
The quantized prediction error signal supplied from the quantized prediction error signal is code-converted using an efficient code such as a Huffman code.

When the control signal indicates encoding of a still image, code conversion is stopped and a code indicating that the data is invalid is output.

Further, the code converter 12 switches between refreshing encoding and frame opening encoding in response to a refresh control signal supplied from the refresh circuit 3. The code-converted signal is matched with the line speed and becomes the output of the code converter 12. The output signal of the code converter 12 is supplied to a moving image decoding device via a high-speed line. When the control signal supplied from the control circuit 4 indicates encoding of a still image, the code converter 13 converts the quantized prediction error signal supplied from the quantizer 7 into a Huffman code or the like. Convert code using efficient code. When the control signal indicates encoding of a moving image, a code indicating that the data is invalid is output. Also, like the code converter 12, it also performs code switching. The code-converted signal is matched with the line speed and becomes the output of the code converter 13.

The output signal of the code converter 13 is supplied to a still image decoding device via a low-speed line.

Next, a still image decoding device will be explained with reference to FIG.

Still images encoded in the refresh mode of the video encoding apparatus are supplied to the code inverter 20 via a line 200 from a low-speed line for still images. The code inverse converter 20 performs inverse code conversion while matching the line speed and decoding speed, and when the region refreshed by intra-frame encoding is reverse code-converted, it performs intra-frame decoding via line 2021. When the area encoded by inter-frame prediction is reversely code-transformed, it is supplied to the adder 22 via line 2o22. Further, the code inverse converter 20 detects from the code whether the signal has been intraframe encoded or the signal has been interframe encoded, and the intraframe decoder 21 and the interframe decomposer A switching signal for selecting between 25 and 25 is supplied to switch 24 via line 2024. The intra-frame decoder 21 performs intra-frame decoding and supplies the value of a decoded pixel near the pixel to be decoded, such as a previous pixel or line, as one signal to the switch 24 . The pre-AI + function at this time is
Use the same one as on the sending side. The adder 22 adds the inverse code-converted interframe signal supplied via the line 2022 and the interframe prediction signal supplied from the frame memory 23 to obtain an interframe decoded signal. At this time, the signal supplied from the code inverse converter 20 is encoded as invalid data between frames during encoding, so the interframe prediction signal supplied from the frame memory 23 is directly output from the adder 22. Become.

The interframe decoded signal output from the adder 22 is
4. The switch 24 switches the intraframe decoder 21,
Alternatively, the output of the adder 22 is selected. The output signal of switch 24 is fed to frame memory 23 and output via line 240, becoming the output of the still image decoding device. The frame memory 23 delays the signal supplied from the switch 24 by one frame time, and
It is supplied to the adder 22 as the lllJ signal.

On the other hand, in the case of a video decoding device, a normal video decoding device having the same prediction function as the video coding device may be used.

Second Embodiment Next, a second embodiment will be described with reference to FIG. 6.

Even if still image encoding is instructed, the video encoding device can treat the signal output from frame memory 1 as a video and continue the normal interframe encoding operation. .

That is, in this case, intra-frame coding is applied to the line to be refreshed and inter-frame coding is applied to the other lines, and the same pre-71!1 error signal is simultaneously supplied to the code converter 12.13. be done. The code converter 12 performs the same operation as in normal video encoding.

The code converter 13 performs normal still image encoding operations. Furthermore, regardless of the control signal for encoding a still image supplied from the control circuit 4 via the line 450, writing to the frame memory 10 only needs to be executed at all times.

In this way, the contents of the frame memories 10 and 23 become the still images supplied from the frame memory 1 when the transmission of the still images is finished. Further, the contents of the frame memory of the interframe decoding loop in the moving image decoding device are also the same as the contents of the frame memory 1.

-Third Example- moreover,! The embodiment shown in FIG. 83 will be explained.

In the above description, a prediction method has been described as intra-frame coding, but this embodiment uses transform coding such as orthogonal transform.

As shown in FIG. 8, an orthogonal transformer 31 transforms the interframe prediction error, a quantizer 32 quantizes the resulting transform coefficients, supplies them to code transformers 12 and 13, and at the same time performs inverse orthogonal transform. It is also supplied to the vessel 33. The inverse orthogonal transformer 33 performs an inverse transform to the orthogonal transform in the orthogonal transformer 31. When the switch 36 is switched, the ground side is selected during intraframe encoding, the ground side is selected during interframe encoding, and the output of the frame memory 35 is selected during interframe encoding.

(Effects of the Invention) As explained in detail above, if the image encoding device of the present invention is used, multi-point communication can be performed simultaneously between a moving image decoding device and a still image decoding device. Conventionally, a still image communication station and a video communication station are conducted in their respective closed areas, and the scope of communication can be greatly expanded.

When the present invention is put to practical use in this way, its effects are extremely large.

[Brief explanation of drawings]

FIG. 1 is a conceptual diagram showing the connection relationship between a moving image decoding device and a still image decoding device according to the present invention, and a moving image encoding device that performs multi-point communication; FIG. b)
, (c) is a conceptual diagram showing refreshing in encoding a moving image according to the present invention, FIG. 3 is a conceptual diagram when encoding a still image by the video encoding device according to the present invention, and FIG. is a block diagram of a moving image encoding device of the present invention, FIG. 5 is a block diagram of a still image decoding device on the receiving side, FIG. 6 is a block diagram of an encoding device according to the first embodiment of the present invention, FIG. 7 is a block diagram of a still image decoding device on the receiving side according to the first embodiment of FIG. 6, and FIG. 8 is a block diagram of a still image decoding device on the receiving side according to the first embodiment of the present invention, which uses transform encoding such as orthogonal transform according to the third embodiment of the present invention. FIG. 1...Frame memory 2...Sync separator, 3.
. . . Refresh circuit, 4 . . Control circuit, 5, 11.
24.36...Switch, 6, 30...Subtractor, 7
, 32... quantizer, 8, 22.34... adder,
9... Frame inner child Δp1 unit, 10, 23.35...
・Frame memory 12.13... code converter, 2
0... Code inverse converter, 21... Intraframe decoder,
31... Orthogonal transformer, 33... Inverse orthogonal transformer.

Claims (1)

[Claims] 1) An image encoding device that uses correlation within a screen and correlation between screens, comprising means (1) for storing one screen of input images, and means (1) for storing the input image and the screen. a first selection means (5) for selecting one of the output signals of the first selection means (5);
means (6) for obtaining a signal with reduced redundancy using the output of 5) and a third prediction signal; means (7) for quantizing the redundancy-reduced signal, which can be set to a predetermined value; and obtaining a locally decoded signal using the third predicted signal and the output of the quantization means (7). means (8); and a first predictor (9) that generates a first prediction signal using intra-screen correlation from the locally decoded signal;
a second predictor (10) capable of generating a second predicted signal using correlation between screens and storing the second predicted signal according to a second instruction; and the first predicted signal. and a second selection means (11) which selects the second prediction signal according to the first instruction and generates the third prediction signal, and converts the code of the output of the quantization means (7), A first code converting means (12) capable of converting the code as invalid regardless of the input according to the second instruction, and converting the code of the output of the quantizing means (7), a second code converting means (13) capable of performing code conversion while invalidating all prediction errors as a result of using the correlation between screens; and a first control means (13) that performs the first instruction.
3), the quantization means (7), and the second predictor (1);
0), a second instruction for giving the second instruction to the first code conversion means (12) and the second code conversion means (13), and a third instruction to the quantization means (7); An image encoding device comprising: a control means (4).