JP3532222B2 - High efficiency coding method and apparatus - Google Patents

Description

Detailed Description of the Invention

[0001]

The present invention relates to a video signal (video signal).
The present invention relates to an encoding device for encoding and a decoding device for decoding, and more particularly to a high efficiency encoding device and a high efficiency decoding method for efficiently encoding and decoding a two-dimensional video signal.

[0002]

2. Description of the Related Art As a high efficiency coding method capable of accurately coding a video signal with a small number of bits, an adaptive dynamic range coding method (hereinafter, referred to as The ADRC method is proposed (see, for example, December 11, 1986, Institute of Electronics and Communication Engineers, paper MR86-43). The processing content will be described. When an image is divided into blocks, local correlations within each block often have a smaller dynamic range. In the ADRC method, as illustrated in FIG. 4, an encoding device (transmission device) divides an image into a plurality of blocks, and a minimum value and a dynamic range (maximum value-minimum value) of each block are divided. calculate. The difference obtained by subtracting the minimum value from each pixel data in each block is calculated, and these differences are adaptively re-encoded using the dynamic range and compressed into data having a smaller number of bits than each pixel data. Since the difference is re-encoded, compressed efficient re-encoded data having a small number of bits can be obtained. The compressed data, the minimum value, and the dynamic range of each pixel are transmitted to the decoding device (reception device) via the transmission system.

As a method of re-encoding a differential video signal using a dynamic range, a fixed length ADRC method and a variable length ADRC method have been proposed (for example, Japanese Unexamined Patent Publication No. 61-61196).
See Japanese Patent Publication No. 147689). The fixed length ADRC method is a method in which, when a new number of bits is assigned for re-encoding, the number of bits is fixed in every block and the quantization step width is changed according to the dynamic range of each block. The variable length ADRC method is a method in which the number of bits is changed for each block according to its dynamic range and the quantization step width is made constant.

FIG. 5 shows a coding apparatus based on the fixed length ADRC method. The input video signal (video signal) SIN is converted by the analog / digital conversion circuit 12 into pixel data in a digital format with a quantization bit number of, for example, 8 bits. The pixel data in digital form is divided into blocks in the blocking circuit 13 for each two-dimensional area of 4 pixels × 4 lines, for example. Pixel data divided into blocks is applied to the minimum value / maximum value detection circuit 14, and the minimum value MIN and the maximum value MAX in each block are detected. In the subtraction circuit 15, the dynamic range DR is calculated by subtracting the minimum value MIN from the maximum value MAX. Pixel data after block partitioning from the blocking circuit 13, the delay circuit 21,
It is delayed by the detection processing time in the minimum / maximum value detection circuit 14 and applied to the subtraction circuit 22, and the minimum value MIN is subtracted from each pixel data to calculate the differential video signal ΔD. The encoding circuit 23 receives the differential video signal ΔD and the dynamic range DR and re-encodes (re-quantizes) the 4-bit pixel data DT by the above-mentioned re-encoding method. The framing circuit 16 uses the minimum value M
The IN, the dynamic range DR, and the re-encoded pixel data DT are frame-processed into a predetermined format and combined into an encoded video output signal SOUT. This encoded video output signal SOUT is output to the decoding device.

In the case of the variable length ADRC method, the coding circuit 2
3, the number of bits of the re-encoded pixel data DT is changed according to the dynamic range DR, and the data indicating the number of bits is transmitted to the decoding device in block units. That is, the high-efficiency coding apparatus based on the fixed length ADRC method and the high-efficiency coding apparatus based on the variable length ADRC method are different only in the processing of the coding circuit 23.

FIG. 6 shows a decoder according to the fixed length ADRC method. The coded video output signal SOUT is applied to the frame decomposing circuit 32 to decompose the framed re-encoded pixel data DT, the minimum value MIN, and the dynamic range DR. The re-encoded pixel data DT and the dynamic range DR are applied to the decoding circuit 33, and the process reverse to that of the encoding circuit 23 is performed to decode the differential video signal ΔD. In the adder circuit 34, the minimum value MI
The decoded difference ΔD is added to N, and the analog / digital conversion circuit 12 in the encoding circuit reproduces 8-bit pixel data equivalent to the converted data. These reproduced pixel data are applied to the decomposition circuit 35, rearranged in the order of the time axis corresponding to the input video signal SIN, and applied to the high-efficiency encoding device shown in FIG. 5 in the digital / analog conversion circuit 36. A signal corresponding to the video signal SIN is decoded.

[0007]

As described above, when the minimum value, the maximum value and the dynamic range are calculated independently for each frame and the differential video signal is re-encoded, for example, the background image or the still image is obtained. In
Even though the image data of consecutive frames originally continues at the same level, the signal level changes due to the influence of noise, etc., and the decoded value after quantization changes slightly,
There is a problem that the encoded image quality is degraded. In particular, when the still image and the image of the background portion fluctuate, the deterioration of the image quality is conspicuous.

A more specific example will be described with reference to FIG. Figure 7
(A) shows the minimum value MIN, the maximum value MAX, and the dynamic range DR of the video signal for one block in a certain frame. It is easy to understand, and as shown in FIG. 7B, an example of quantization with 1 bit of "0" or "1" is shown. What is indicated by a small circle indicates image data that has not changed originally. The arrow indicates the quantization reference value. In this example, in the first frame, the minimum value MIN has risen somewhat due to noise, but since the pixel data exceeds the quantization reference value, the pixel data is quantized as "1". In the second frame, the minimum value MIN is further lowered, and therefore the pixel data is quantized as “1”. In the third frame, the minimum value MIN rises due to noise and the pixel data falls below the quantization reference value, so that the pixel data is quantized as “0”. In the fourth frame, the minimum value MIN has decreased, so the pixel data is quantized as "1". Thus, dynamic by the signal level constant pixel data due to the influence of noise or the like at a critical value in the vicinity of the quantization reference value varies
The range has changed, and it changes as originally indicated by a small circle.
The quantized value of non- existing image data becomes "1" or "0"
It becomes and fluctuates.

Such a variation in the quantization result causes an error in re-encoding, and finally causes deterioration in image quality of the video signal decoded and reproduced by the receiving device.

Therefore, an object of the present invention is to provide a high-efficiency coding method and its apparatus that enable accurate coding without depending on the above-mentioned quantization fluctuation between frames. It is another object of the present invention to provide a high efficiency decoding method corresponding to the above high efficiency encoding method and a high efficiency decoding device corresponding to the above high efficiency encoding device.

[0011]

As described above, generally, in the still image portion such as the background, the image data has noise,
The maximum value and the minimum value fluctuate with time, and when ordinary encoding processing and decoding processing are performed, the decoding result also fluctuates with the passage of time. In order to avoid the fluctuation with the passage of time,
In the still image section, the dynamic range is calculated from two consecutive frames, and the fluctuation with time is absorbed. Therefore, in the present invention, the presence or absence of an image change is detected for at least two consecutive frames, and when it is considered to be a still image or a background portion, a new dynamic range is calculated from the dynamic range of consecutive frames to obtain pixel data. For this, a high efficiency encoding process such as the ADRC method is performed using this dynamic range. Further, in the present invention, a decoding process corresponding to this encoding is performed.

According to the first aspect of the present invention, the maximum value of the pixel data and the minimum value of the pixel data included in each of the predetermined regions formed by a plurality of pixels in the frame of the input video signal. A maximum value / minimum value detecting means for detecting the maximum value and the minimum value in each predetermined area.
Dynamic range calculation means for calculating the dynamic range of each area and the dynamic range of each predetermined area
As the dynamic range within each area of the previous frame
The storing means that has been stored and the previous storing means
The dynamic range of each predetermined area of the frame
Various parts of the current frame calculated by the dynamic range calculation means
Average the dynamic range of the constant area and
A new dynamic range calculation means for calculating as a new dynamic range, a dynamic range of each predetermined area calculated by the dynamic range calculation means, and a new dynamic range of each predetermined area calculated by the new dynamic range calculation means Input , and according to the switching command
Enter the dynamic range or the new dyna
Switching means for outputting any one of the range, an image of each predetermined area of the current frame, and the front frame.
It is determined whether the image change between the image of each predetermined area of the frame is large or small, and when the image change is large, it is calculated by the dynamic range calculation means from the switching means .
Switching to output the dynamic range of each predetermined area
Command is output, and when the image change is small, a new dynamic range of each of the predetermined areas is output from the switching means.
Output a switching command to output , the determination means, based on the number of quantization bits determined for the predetermined region, by dividing the dynamic range or the new dynamic range output from the switching means. set the resulting level range, and an encoding means that the value obtained by subtracting the minimum value from the value of the pixel data to generate an encoded signal corresponding to the level range belongs, the
The storage means calculates by the dynamic range calculation means.
The dynamic range of each predetermined area of the current frame.
Stored as the dynamic range of each predetermined area of the frame
An encoding device is provided.

The determination means compares the minimum value of the pixel data of the current frame with the minimum value of the pixel data of the previous frame, and the dynamic range of the current frame and the dynamic range of the previous frame. It is characterized in that it is determined that the image change is small when it is within a predetermined range. Further, it is characterized in that it has block forming means for dividing the inputted video signal into blocks of a predetermined size, and the divided blocks are set as the predetermined area.

According to the second aspect of the present invention, the maximum value of the pixel data and the minimum value of the pixel data included in each of the predetermined regions formed by a plurality of pixels in the frame of the input video signal. the maximum and minimum values detecting step of detecting, each place from the maximum value and the minimum value in said each predetermined area
A dynamic range calculating step of calculating a dynamic range of the constant region, respectively, stored in the storage means
The dynamic range of each predetermined area of the previous frame, and
Current frame calculated in the dynamic range calculation step
The dynamic range of each predetermined area of
A new dynamic range calculation step of calculating as a new dynamic range of the area, a dynamic range of each predetermined area calculated in the dynamic range calculation step, and a new dynamic range of each predetermined area calculated in the new dynamic range calculation step are input. ,
Input according to the switching command,
Or output one of the new dynamic ranges
That a switching step, the predetermined said current frame
Between the image of the area and the image of each predetermined area of the previous frame
Determining whether the image change is large small image change is large
When the threshold is reached, the die
Before each predetermined area calculated in the Namic range calculation step
Outputs a switching command to output the dynamic range
However, when the image change is small , the new dynamic range of each of the predetermined areas is output in the switching step.
Output a switching command to force the determination step and the dynamic range output in the switching step or the new dynamic range obtained by dividing based on the number of quantization bits determined for the predetermined region. An encoding step of generating an encoded signal corresponding to the level range to which a value obtained by subtracting the minimum value from the value of the pixel data belongs by setting the level range set , and the current range calculated in the dynamic range calculating step.
Pre-frame the dynamic range of each given area of the frame
Comprising the step of storing in the storage means as a dynamic range of each predetermined area of the beam, the encoding method is provided.

The high-efficiency decoding process in the present invention and the device can be applied to existing ones.

[0016]

As described above, generally, in the still image portion such as the background, the image data has noise, and the maximum value and the minimum value are time.
It changes with the passage of time, and when ordinary encoding processing and decoding processing are performed, the decoding result also changes with the passage of time. In order to avoid the variation with the passage of time, the dynamic range is calculated from two consecutive frames in the background portion and the still image portion, and the variation with the passage of time is absorbed.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a flowchart showing an encoding process as an embodiment of the high efficiency encoding method of the present invention. Step 1 : Blocking process The 2D video signal is divided into a plurality of blocks having a predetermined size. Such blocking processing is performed in the present invention.
Although not essential, it is preferable because the signal processing described later, particularly the background portion or the still image portion can be performed for each block.

Step 2 : Calculation of minimum value and dynamic range For each of the divided blocks, the pixel data in that block is examined to determine the minimum value MIN and maximum value MAX.
And the dynamic range DR (= maximum value MAX-minimum value MIN) is calculated. The calculated minimum value MIN and dynamic range DR are stored for calculating the dynamic range of the next frame. Steps 3 and 4 : Detection of presence or absence of image change between previous frame and current frame The difference between the minimum value MIN _-1 of the corresponding block in the previous frame and the minimum value MIN of the block of the current frame is the threshold value TH1. Determine whether it is within. When it is within the threshold value TH1, it is further determined whether or not the difference between the dynamic range DR ₋₁ for the block in the previous frame and the dynamic range DR for the block in the current frame is within the threshold value TH2. . That is, in order to detect whether it is the still image portion or the background portion based on the presence or absence of the change in the image data of the previous frame and the current frame, the variation of the minimum value MIN and the variation of the dynamic range DR are determined.

Step 5 : Calculation of new dynamic range for two frames For consecutive frames, if the image of the corresponding block can be treated as a still image or a background, it is consecutively calculated .
A new dynamic range DR is calculated from the dynamic range of two frames. As one of the methods for calculating the new dynamic range DR over the two frames, a simple average value of the dynamic range DR ₋₁ of the previous frame and the dynamic range DR of the current frame is used. Alternatively, as another method of calculating the new dynamic range DR, a large weight is attached to the current dynamic range DR, for example, 0.6, and a small weight is attached to the previous dynamic range DR ₋₁ , for example, 0.4. It is also possible to perform a weighted averaging process to obtain the average.

[0020] Step 6: New for the differential video signal ΔD obtained by subtracting the minimum value MIN from the block of pixel data for a new dynamic range ADRC processing This time frame using the
Re-encoding processing by the ADRC method is performed using the dynamic range DR . As the re-encoding process by the ADRC method, the fixed length ADRC method or the variable length ADR described above is used.
Either method C may be applied. Step 8 : Minimum value MIN, new dynamic range D
The data obtained by re-encoding R and the differential video signal ΔD is framed.

Steps 7 and 8 : The image data of the current frame greatly changes from the image data of the previous frame,
For the blocks that are not recognized as still images or background blocks, the dynamic range DR of this frame
Is used to perform a normal re-encoding process and a framing process.

FIG. 2 is a diagram showing the quantization processing according to the above embodiment corresponding to FIG. 2A shows the minimum value MI.
The relationship between N, the maximum value MAX, and the dynamic range DR is shown. FIG. 2 (B) shows the quantization result in the first to fourth frames assuming 1-bit quantization. The small circle is the original image signal value, and the arrow is the quantized value. The fluctuation of the minimum value MIN is the same as that in FIG. Continuous previous frame
2 frames of dynamic range with the frame and the current frame
For example, a dynamic range of 2 frames
Or simply averaging the two
By calculating the weighted average for the range, calculating the new dynamic range, and quantizing the image data using this new dynamic range, the third frame is quantized as “0” due to noise in the example of FIG. 7B. However, it is quantized as “1” in FIG. Also,
The second and fourth frames are also quantized as "1" with a sufficient margin.

As described above, according to the present embodiment, it is possible to suppress the time variation due to noise in the background portion and the like, and to realize the encoding that does not adversely affect the scene change and the like.

FIG. 3 is a block diagram of a high-efficiency coding apparatus as one embodiment for realizing the processing contents illustrated in FIG. This circuit configuration is an improved circuit of the high efficiency coding device illustrated in FIG. The high-efficiency encoder of FIG.
Digital conversion circuit 12, blocking circuit 13, minimum value / dynamic range detection circuit 14A, memory 17, determination circuit 18, new dynamic range calculation circuit 19, switching circuit 20, delay circuit 21, subtraction circuit 22, encoding circuit 23, It has a framing circuit 16.

The analog / digital conversion circuit 12 converts the input video signal SIN into digital pixel data. The blocking circuit 13 blocks the digital two-dimensional pixel data into a size of, for example, 8 × 8 pixels. The size of this block formation should be such that the background portion can be detected appropriately. The minimum value / dynamic range detection circuit 14A is the minimum value / maximum value detection circuit 1 in FIG.
4 and the subtraction circuit 15 are integrated, and the minimum value MIN from the pixel data in each block is
Maximum value MAX is calculated, and dynamic range DR
(= Maximum value MAX-minimum value MIN) is calculated. The memory 17 uses the minimum value and
Minimum value M calculated by the dynamic range detection circuit 14A
Save IN and dynamic range DR.

The judgment circuit 18 makes the judgments shown in steps 3 and 4 of FIG. When the determination result of the determination circuit 18 is affirmative and the image data of the block can be treated as the background, the determination circuit 18 activates the new dynamic range calculation circuit 19 and the new dynamic range calculation circuit 19 operates for 2 frames. A new dynamic range DR over the range is calculated by the method described above. As a result, the new dynamic range calculation circuit 19
Is the simple average value of the dynamic range of 2 frames or
It is calculated as a new dynamic range if it is a multiple average value.
The dynamic range DR of the current frame calculated by the minimum value / dynamic range detection circuit 14A and the new dynamic range DR from the new dynamic range calculation circuit 19 are input to the switching circuit 20, and the determination circuit 18 makes the above determination. Thus, when the image data of the block is considered as the background image, the new dynamic range DR is applied to the blocking circuit 13, and otherwise, the dynamic range DR from the minimum value / dynamic range detection circuit 14A is encoded. The switching circuit 20 is selectively operated so as to be applied to the circuit 23.

The delay circuit 21 delays the pixel data from the blocking circuit 13 by a time corresponding to the processing time of the minimum value / dynamic range detection circuit 14A. That is,
The timing is matched with the calculation time in the minimum value / dynamic range detection circuit 14A. The subtraction circuit 22
The minimum value M calculated by the minimum value / dynamic range detection circuit 14A from the pixel data output from the delay circuit 21.
The difference video signal ΔD is calculated by subtracting IN. The encoding circuit 23 outputs the differential video signal ΔD to the switching circuit 2
New dynamic range DR selectively output from 0
Alternatively, re-encoding is performed using the dynamic range DR of the current frame. The framing circuit 16 uses the data DT re-encoded by the encoding circuit 23 and the switching circuit 2
The dynamic range DR output from 0 and the minimum value MIN are framed and output as a coded video output signal SOUT to the transmission path.

The encoded video output signal SOUT is received by the high-efficiency decoding device via the transmission path and is decoded.
As the high efficiency decoding device, the high efficiency decoding device illustrated in FIG. 6 can be used. Since the high-efficiency decoding device performs decoding processing on the received minimum value MIN, dynamic range DR, and data obtained by re-encoding the differential video signal ΔD, the received dynamic range DR
Is the new dynamic range DR or the dynamic range DR of the current frame, the decoding process is performed. Therefore, in this embodiment, the high efficiency decoding device illustrated in FIG. 6 can be used. In other words,
The present invention has an advantage that it is not necessary to newly manufacture a high-efficiency decoding device.

[0029]

According to the present invention, with high efficiency,
It is possible to provide a high-efficiency coding method and its apparatus which have a small number of coding bits and are not affected by fluctuations caused by noise over time . Further, according to the present invention, the high-efficiency decoding method corresponding to the high-efficiency encoding method may be an existing method, and the existing high-efficiency decoding apparatus corresponding to the high-efficiency encoding apparatus may be applied. it can.

[Brief description of drawings]

FIG. 1 is a block diagram showing an encoding process of a high efficiency encoding method of the present invention.

FIG. 2 is a graph illustrating the processing result shown in FIG.

FIG. 3 is a circuit configuration diagram of a high efficiency encoding device of the present invention.

FIG. 4 is a graph illustrating a basic encoding process of the high-efficiency encoder of the present invention.

[Fig. 5] Fig. 5 is a configuration diagram of a high-efficiency encoding device which is a basis of the high-efficiency encoding device of the present invention.

6 is a circuit configuration diagram of a high-efficiency decoding device corresponding to the high-efficiency encoding device shown in FIG.

7 is a diagram showing a process in the high efficiency encoding device shown in FIG.

[Explanation of symbols]

12-Analog / digital conversion circuit 13-Blocking circuit 14A-Minimum value-Dynamic range detection circuit 15-Subtraction circuit 16-Frame conversion circuit 17-Memory 18-Determination circuit 19-New dynamic range Calculation circuit 20 Switching circuit 21 Delay circuit 22 Subtraction circuit 23 Encoding circuit SIN Input video signal SOUT Encoded video output signal 32 Frame decomposing circuit 33 Decoding circuit 34 ..Addition circuit 35..Block decomposition circuit 36..Digital / analog conversion circuit

Claims (6)

(57) [Claims] 1. A maximum value / minimum value detection for detecting a maximum value of pixel data and a minimum value of the pixel data included in each of a predetermined area formed by a plurality of pixels in a frame of an input video signal. Means and within each predetermined area from the maximum and minimum values in each predetermined area
A dynamic range calculating means for calculating a dynamic range of each, each region of the previous frame the dynamic range of the predetermined area
Means of storing as the dynamic range in
And a predetermined area of the previous frame stored in the storage means.
Dynamic range and dynamic range calculator
The dynamic level of each predetermined area of the current frame calculated by
And a new dynamic range for each predetermined area
And the new dynamic range calculating means for calculating a respective predetermined area calculated by the dynamic range calculating means
Of the dynamic range and the new dynamic range of each predetermined area calculated by the new dynamic range calculation means, and input in response to a switching command.
Range or the new dynamic range
Output means , a switching means, an image of each predetermined area of the current frame and the previous frame
It is determined whether the image change between the image of each predetermined area is large or small, and when the image change is large , each position calculated by the dynamic range calculation means from the switching means
Switching finger that outputs the dynamic range of the constant region
When the image change is small, the switching means outputs a new dynamic range of each of the predetermined areas.
Determining means for outputting a switching command to be applied , and the dynamic range output from the switching means or the new dynamic range obtained by dividing based on the number of quantization bits determined for the predetermined region. set the resulting level range, and an encoding means for generating an encoded signal corresponding to the level range minus belongs to said minimum value from the value of the pixel data, the storage means, the Calculated by dynamic range calculation means
The dynamic range of each given area of the current frame
Described as the dynamic range of each predetermined area of the previous frame
An encoding device that remembers . 2. The determination means compares the minimum value of the pixel data of the current frame with the minimum value of the pixel data of the previous frame, and the dynamic range of the current frame and the dynamic range of the previous frame. The encoding device according to claim 1, wherein it is determined that the image change is small when the difference is within a predetermined range. 3. has a blocking means for dividing the video signal the input to a predetermined size of the block, the segment block, characterized in that the predetermined area, according to claim 1 or 2. The encoding device according to item 2 . 4. A maximum value / minimum value detection for detecting a maximum value of pixel data and a minimum value of the pixel data included in each of a predetermined area formed by a plurality of pixels in a frame of an input video signal. Step, and within each predetermined area from the maximum and minimum values in each predetermined area
A dynamic range calculating step of calculating a dynamic range of each, Dyna each predetermined region before stored in the storage unit frame
Mick range and the dynamic range calculation step
Dynamic range of each predetermined area of the current frame calculated in
Smoothed out and di Rights and the new dynamic range of the predetermined area
And a new dynamic range calculation step, and each predetermined value calculated in the dynamic range calculation step.
Enter the new dynamic range for each predetermined area calculated by the dynamic range and the new dynamic range calculating step region, said type depending on the switching command
Dynamic range or the new dynamic range
A switching step of outputting one of the following, the image of each predetermined area of the current frame and the previous frame.
Determining whether the image change is large or small between the image of the predetermined area, when the image change is large in the dynamic range calculating step in said switching step
Output the calculated dynamic range of each predetermined area.
Output a switching command to cause a new image to change in each of the predetermined areas when the image change is small.
Output a switching command to output the dynamic range ,
And a level range obtained by dividing the dynamic range or the new dynamic range output in the switching step, based on the determination step and the number of quantization bits determined for the predetermined area. , A coding step of generating a coded signal corresponding to the level range to which a value obtained by subtracting the minimum value from the value of the pixel data belongs , and a current frame calculated in the dynamic range calculation step.
Dynamic range of each predetermined area of the
In the storage means as the dynamic range of each predetermined area
Comprising the steps of: storing, and encoding method. 5. In the determining step, the minimum value of the pixel data of the current frame and the minimum value of the pixel data of the previous frame, and the dynamic range of the current frame and the dynamic range of the previous frame are compared, 5. The encoding method according to claim 4 , wherein it is determined that the image change is small when the difference between is within a predetermined range. 6. further comprising a block step of partitioning the block of the input video signal of a predetermined size, characterized in that the partitioned block and the predetermined area, according to claim 5, wherein Encoding method.