JP3581382B2 - Encoding device and method - Google Patents

Description

[0001]
[Industrial applications]
The present invention relates to a coding apparatus and a coding method, and more particularly to a coding apparatus and a coding method for coding image data by adaptively using an intra-screen coding mode and an inter-screen coding mode.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, when digitally transmitting image information, various encoding methods have been proposed to reduce the amount of transmitted data.
[0003]
As one of the proposed coding methods, there is a method of performing coding by switching between intra-frame coding and inter-frame coding.
[0004]
Intra-frame compression is a method of reducing the amount of information by using characteristics of an image in which adjacent pixels have similar brightness and color.
[0005]
In an actual image, almost the same brightness and color continue almost in the sky and the walls such as walls, so that only 1/5 to 1/10 can be compressed by using only intra-frame compression.
[0006]
The inter-frame compression is a method of obtaining an image using only similar information by using similar images.
[0007]
In a normal moving image, the patterns of adjacent frames are similar, although there are some movements and deformations. Utilizing this point, first, the similarity (motion, color, brightness, etc.) between the frame to be compression-encoded and the adjacent frame is calculated. Based on the calculation, a value of a frame that is more similar to the “predicted value”, that is, the “frame to be encoded” from the “proximal frame” is calculated.
[0008]
Next, only the difference information from the "prediction value" is encoded (recorded / transmitted) from the frame to be encoded. Therefore, the data amount (correction amount) is reduced.
[0009]
In other words, when a person moves to the right in a moving image of a person, the pixel in which the person including the movement correction information is included in the previous frame is the predicted value, and the predicted value is calculated from the entire pixels moved to the right. The difference is the difference.
[0010]
In the case of a conventional encoding device, it is generally known that when an error occurs on a transmission path when compression is performed by inter-frame processing, the error propagates. Therefore, when a predetermined number of inter-frame processes are performed, an intra-frame process (refresh operation) is automatically performed.
[0011]
[Problems to be solved by the invention]
The amount of data per frame of each encoded screen is about 16K to 25K bytes for intra-frame processing and about 7 to 10K bytes for inter-frame processing.
[0012]
That is, it is generally known that the intra-frame processing generates a larger amount of data than the inter-frame processing.
[0013]
Therefore, when performing intra-frame coding periodically, if the quantization is performed using the immediately preceding quantization step (the quantization step used for inter-frame coding), the amount of generated data suddenly increases, and the transmission rate increases. This is a problem.
[0014]
Therefore, if encoding is performed by changing the quantization step in order to suppress the amount of generated data, there arises a problem that image quality is deteriorated this time.
[0015]
The above-described problem occurs in a coding apparatus (having at least a predictive coding mode) that performs coding by adaptively switching a plurality of coding modes.
[0016]
In view of the above-described background, an object of the present invention is to solve the above-described problems in a conventional encoding device and to provide an encoding device and a method thereof that prevent signal degradation.
[0017]
[Means for Solving the Problems]
For this purpose, the present application provides an input device for continuously inputting image data and an input device for selectively inputting an intra-screen encoding mode and an inter-screen encoding mode. Encoding means for encoding the image data inputted by the means; and dividing the image data inputted by the input means into a plurality of areas, and a desired code amount when the image data is encoded for each area. Code amount control means for controlling the encoding means so that the code amount is equal to, and forcibly executing the intra-screen encoding mode at a constant ratio in each area which is a control unit of the code amount control means. Refresh control means for setting a refresh area and controlling the refresh area to be moved in predetermined units on a screen-by-screen basis. In addition, the present application, as an encoding method of the present invention for such purpose, an input step of continuously inputting image data, and selectively using an intra-screen encoding mode and an inter-screen encoding mode, An encoding step of encoding the image data input in the input step, and a code amount when dividing the image data input in the input step into a plurality of areas and encoding the image data for each area Is a code amount control step of controlling the encoding means so that the desired code amount is obtained, and the intra-screen encoding mode is forcibly set at a fixed ratio in each area which is a control unit of the code amount control step. And a refresh control step of setting a refresh area to be executed and controlling the refresh area to be moved in predetermined units on a screen basis.
[0019]
【Example】
Hereinafter, an encoding device according to an embodiment of the present invention will be described.
[0020]
FIG. 1 is a block diagram of an encoding device to which the present embodiment is applied.
[0021]
In FIG. 1, reference numeral 101 denotes an input terminal to which an image signal is input; 102, an A / D conversion circuit that converts an input analog signal into a digital signal; 103, a block division circuit that divides the order of pixels into blocks; Reference numeral denotes a prediction error calculation circuit that performs a difference operation between the current frame signal and the prediction signal, and reference numeral 105 denotes a switch circuit that switches between the input from the block divider 103 and the input from the prediction error calculation circuit 104.
[0022]
A determination circuit 106 compares the current frame signal with the prediction error signal and selects an encoded signal.
[0023]
107 is a refresh control circuit. To the refresh control circuit 107, a frame signal and W and S setting values to be described later are input. The operation of the refresh control circuit 107 will be described later in detail.
[0024]
Reference numeral 108 denotes an OR circuit that receives the determination result of the determination circuit 106 and the refresh control signal from the refresh control circuit 107, 109 denotes an orthogonal transform circuit that orthogonally transforms the signal selected by the switch circuit 105, and 110 denotes an orthogonal transform coefficient. A quantizing circuit for quantizing, 111 is a variable length coding circuit for performing variable length coding of the coefficient quantized by the quantizing circuit 110, 112 is a buffer memory for controlling the amount of data generated and the amount of data transmitted , 113 are transmission I / F circuits for creating a transmission format, and 114 is a transmission signal output terminal.
[0025]
Reference numeral 115 denotes an inverse quantization circuit that performs inverse quantization of the quantization circuit 110; 116, an inverse orthogonal transformation circuit that performs inverse transformation of the orthogonal transformation circuit 109; 117, an addition circuit for reproducing an image of the current frame; A switch circuit that selects the terminals a and b based on the output of the OR circuit 108 in conjunction with the circuit 105.
[0026]
An image memory circuit 119 temporarily stores a reproduced image of the current frame, a motion vector detecting circuit 120 compares a signal of the current frame with a signal of the previous frame to calculate a motion vector, and a reference numeral 121 denotes a motion vector for the reproduced image signal of the previous frame. This is a motion compensation circuit that performs motion compensation according to the signal of the vector detection circuit 120. Reference numeral 130 denotes a buffer control circuit for keeping the amount of data stored in the buffer memory 112 constant.
[0027]
Hereinafter, the operation of the encoding device having the above configuration will be described.
[0028]
The analog image signal input to the input terminal 101 is converted into a digital signal by an A / D conversion circuit 102, and further divided by a block divider 103 into blocks of a pixel in the horizontal direction and b lines in the vertical direction, for example, a = b = 8 Is done.
[0029]
This signal is input to the prediction error calculating circuit 104, the end of the intra-frame / inter-frame encoding changeover switch circuit 105a, and the intra-frame / inter-frame encoding determining circuit 106.
[0030]
The prediction error calculation circuit 104 receives the motion prediction value of the previous frame and the signal of the current frame, calculates the difference between them, and inputs the difference value to the end of the switch circuit 105b and the determination circuit 106.
[0031]
The determination circuit 106 compares the input current frame signal and the prediction error signal, and determines the one with the higher coding efficiency (inter-frame / intra-frame coding) as a determination output.
[0032]
The decision output controls the switch circuits 105 and 118 via the OR circuit 108, and the terminal a is adaptively selected if the current frame has better coding efficiency, and the terminal b is suitable if the prediction error signal is better.
[0033]
To the other of the OR circuits 108, a control signal from the refresh control circuit 107 for periodically setting a forced frame for refresh is input.
[0034]
The purpose of this refresh operation is to stop the propagation of an error when an error occurs in the image.
[0035]
Details of the operation of the refresh control circuit 107 will be described later.
[0036]
The signal selected by the switch circuit 105 is input to the orthogonal transform circuit 109.
[0037]
In the present embodiment, a discrete cosine transform (DCT) which has high conversion efficiency and can be realized by hardware is applied as an orthogonal transform method.
[0038]
The transformed coefficients are output from the orthogonal transformation circuit 109 and input to the quantization circuit 110.
[0039]
The quantization circuit 110 quantizes the input orthogonal transform coefficients according to characteristics in consideration of the visual characteristics and the like of the image. The output signal of the quantization circuit 110 is input to the variable length code circuit 111 and the inverse quantization circuit 115, respectively.
[0040]
The variable-length coding circuit 111 rearranges the data after the orthogonal transformation so that the data generation distribution is efficiently coded from the characteristics of the image, and the characteristics of the distribution are optimized for the coding efficiency. Such variable length coding is performed.
[0041]
The output signal from the variable length coding circuit 111 is controlled by the buffer control circuit 130 so that the data amount of the buffer memory input to the buffer memory 112 is constant in a predetermined processing unit (described later in detail). Assuming that the buffer amount has increased, a control signal for selecting a quantization table for suppressing data generation is sent from the buffer control circuit 130 to the quantization circuit 110 and the inverse quantization circuit 116. Conversely, when the buffer amount decreases, a control signal for selecting a quantization table in which the data generation amount increases is sent. By the processing operation, the amount of data stored in the buffer memory 112 is kept constant. Further, the operation is performed so that the output data amount from the buffer memory 112 becomes equal in unit time. The output from the buffer memory 112 is input to the transmission I / F circuit.
[0042]
The transmission I / F circuit 113 outputs the image data in the buffer memory 112 to the output terminal 114 in accordance with the transmission format in synchronization with the transmission clock. At this time, motion vector information, a transmission synchronization signal, an error correction code, and the like included in the transmission format are also multiplexed and transmitted.
[0043]
On the other hand, the signal input to the inverse quantization circuit 115 is the same signal as the input signal of the quantization circuit 110 due to the opposite characteristics of the quantization circuit 110 and is output from the inverse quantization circuit 115 and input to the inverse orthogonal transform circuit 116. Is done. Similarly, an inverse orthogonal transform circuit 116, which is an inverse transform circuit of the orthogonal transform circuit 109, outputs the same signal as the input signal of the orthogonal transform circuit 109.
[0044]
The inversely converted signal and the output signal from the switch 118 are input to the addition circuit 117.
[0045]
When the switch circuit 118 is connected to the terminal a, the processing in the frame is performed, and the terminal a is similarly selected for the switch circuit 105. The signal of the current frame is coded and inverse coded without taking a difference, and the addition circuit 117 is used. Is input to
[0046]
At this time, since the terminal a of the switch circuit 118 is “0”, the output of the adder circuit 117 becomes equal to the output signal of the block dividing circuit 103.
[0047]
When the switch circuit 118 is at the end b, the output of the motion compensation circuit 121, that is, the difference from the predicted value is output from the switch circuit 105 and input to the addition circuit 117.
[0048]
Since the terminal b of the switch circuit 118 is the same signal as the predicted value, the output is similarly made equal to the output of the block dividing circuit 103 by being added by the adding circuit 117.
[0049]
The image memory 119 stores the current frame signal and outputs it with a delay of about one frame.
[0050]
The motion vector detection circuit 120 receives the output of the image memory 119, that is, the image data of the current frame, compares it with the image signal of the previous frame stored in the image memory 119, and determines the motion of the encoded block as a motion vector. The calculated value is output to the motion compensation circuit 121.
[0051]
The motion compensation circuit 121 performs motion compensation on the image signal of the previous frame based on the vector information from the motion vector detection circuit 120, and outputs a predicted value.
[0052]
Hereinafter, the control operation of the refresh control circuit 107 will be described in detail.
[0053]
First, the relationship between the refresh control signal and the buffer control will be described with reference to FIGS.
[0054]
FIG. 2A shows the processing unit of the buffer control. First, the buffer control circuit 130 detects the amount of data generation in the area corresponding to (a) of FIG. 2A, performs an error calculation on the control target value, and sends the error calculation to the quantization circuit 110 and the inverse quantization circuit 116. A control signal is sent to control the amount of data generated. This operation is also performed sequentially for (a) and (c) in FIG. In each of the areas (A), (A), and (C) in FIG. 2A, each area is controlled by the same control value, so that the units of (A), (A), and (C) are This is a unit of processing for buffer control.
[0055]
The smaller this processing unit, the faster the response of buffer control and the better the response of buffer control, but on the other hand, the amount of data generated is uniform for each small area on the screen, and the data amount is originally small The portion having a large value is greatly compressed, and the image quality is greatly deteriorated. On the other hand, a simple image is assigned a sufficient data amount, and the image is less deteriorated. This is visually noticeable and significantly reduces image quality.
[0056]
Conversely, when the processing unit is increased, the image quality of the encoded data is improved, but buffer control becomes difficult, and a large buffer amount is required. Therefore, the processing unit needs to keep the size of a certain condition. Generally, it is set to an integral multiple of the DCT processing unit.
[0057]
A signal (FIG. 3B) indicating the above-described processing unit is transmitted from the buffer control circuit 130 to the refresh control circuit 107 in FIG.
[0058]
The refresh control circuit 107 operates to output a fixed amount of refresh control signal for each buffer control processing unit (FIG. 3C).
[0059]
FIG. 2B shows an example of the refresh area.
[0060]
Since the refresh area is set at a fixed ratio for each buffer control processing unit, there is no sudden change in data generation due to forced intra processing due to refresh, stable buffer control can be performed, and rapid image quality deterioration can be avoided. Can be. Further, the refresh area is moved by a predetermined unit in a frame unit as shown in FIG. In other words, the generation period of the refresh control signal in FIG. 3C is shifted by a predetermined unit in a frame unit and makes one cycle in n frames.
[0061]
In the present embodiment, the description has been given of the inter-frame motion compensation coding, but the same applies to the inter-field motion compensation coding. Further, in FIG. 3C, the generation timing of the refresh control signal is the same for each buffer control unit ((A), (A), and (C) in FIG. 3A), but is different. It is self-evident.
[0062]
In addition, since the first image has no reference image, it is clear that the intra process is performed.
[0063]
【The invention's effect】
As described above, according to the present invention, a refresh area is set at a fixed ratio in each area which is a code amount control unit, and the refresh area is moved by a predetermined unit in a screen unit. Propagation can be stopped in units of a predetermined number of screens, and the effect of an increase in the data amount due to the refresh area is the same in each area as the code amount control unit, so that a difference in image quality between areas as the code amount control unit is prevented. The overall image quality can be greatly improved.
[Brief description of the drawings]
FIG. 1 is a block diagram of an encoding device according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining a buffer control operation and a refresh operation according to the embodiment of the present invention.
FIG. 3 is a diagram for explaining a generation period of a refresh control signal according to the embodiment of the present invention.
[Explanation of symbols]
105, 118 switch circuit 106 determination circuit 107 refresh circuit 108 OR circuit 130 buffer control circuit

Claims (2)

Input means for continuously inputting image data;
Encoding means for encoding the image data input by the input means, by selectively using the intra-screen encoding mode and the inter-screen encoding mode,
A code amount for dividing the image data input by the input unit into a plurality of areas, and controlling the encoding unit so that a code amount when the image data is encoded for each area becomes a desired code amount. Control means;
A refresh area for forcibly executing the intra-screen encoding mode is set at a fixed ratio in each area, which is a control unit of the code amount control means, and the refresh area is moved in predetermined units in screen units. And a refresh control means for performing control as described above. An input step of continuously inputting image data;
An encoding step of encoding the image data input in the input step by selectively using an intra-screen encoding mode and an inter-screen encoding mode,
A code amount that divides the image data input in the input step into a plurality of areas, and controls the coding unit so that a code amount when the image data is coded for each area is a desired code amount. Control process;
A refresh area for forcibly executing an intra-screen encoding mode is set at a fixed ratio in each area, which is a control unit of the code amount control step, and the refresh area is moved in predetermined units in screen units. And a refresh control step of performing control as described above.

Images