KR100203708B1 - Improved image coding system having functions for controlling generated amount of coded bit stream - Google Patents

Abstract

According to the present invention, in a coded system including MC-DCT and quantization, a complexity of an image to be currently encoded is calculated based on a quantization error value between a current frame input for encoding and a reconstructed previous frame, and the calculation is performed. The present invention relates to an image encoding system having a bit generation amount adjusting function for adaptively adjusting a bit generation amount after encoding by selectively removing high frequency components of an input video signal using a one-dimensional low pass filter. The present invention calculates a quantization error value for each frame based on a pixel value difference signal between a current frame and a reconstructed previous frame, averages the calculated quantization error values of the corresponding frame, and averages the quantization error for a plurality of preset frames. Quantization error calculation means for calculating a value; Refer to the calculated average error value of the plurality of frames as the complexity of the frame to be currently encoded by the encoding means, and for limiting the frequency passband bandwidth of the current frame input for encoding based on the average error value of each detected frame. Control means for generating a filter control signal; And based on the generated filter control signal, the input current frame is provided to the encoding means as a current frame signal for motion estimation and compensation, or the one-dimensional low pass filtering is applied to the input current frame to limit the pass band. By including the one-dimensional low-pass filtering means for providing the encoding means to the encoding means as a current frame signal for motion estimation and compensation, the amount of bits generated after encoding in the encoding means without increasing the excessive step size effectively It is adjustable.

Description

Image Coding System with Bit Rate Control

1 is a block diagram of a video encoding system having a bit generation amount adjusting function according to a preferred embodiment of the present invention.

2 is an illustration of an 8x8 pixel block as an example in accordance with the present invention.

3 is an exemplary view showing a one-dimensional lowpass filter coefficient of order 7 as an example according to the present invention.

4 is an exemplary view illustrating a horizontal filtering process at (0,4) and a vertical filtering process at (3,0) as an example according to the present invention.

* Explanation of symbols for main parts of the drawings

100170: frame memory 110: subtractor

120: image coding block 130: entropy coding block

140: transmission buffer 150: video decoding block

160: adder 180: current frame prediction block

210: quantization error calculation block 220: filter control block

230: low pass filtering block

The present invention relates to a video encoding system for compressing and encoding a video signal. More specifically, the present invention relates to a video encoding system that compresses and encodes a video signal using a motion compensation differential pulse code modulation (MC-DPCM) technique. The present invention relates to an image encoding system having a bit generation amount adjustment function suitable for adaptively adjusting a generation bit amount after encoding with reference to a variation of an input image signal predicted based on a quantization error value between previous frames.

As is well known in the art, the transmission of discrete video signals can maintain better image quality than analog signals. When a video signal consisting of a series of image frames is represented in digital form, a significant amount of data must be transmitted, especially in the case of high quality television (aka HDTV). However, since the usable frequency range of the conventional transmission channel is limited, in order to transmit a large amount of digital data, it is necessary to compress the transmitted data and reduce the amount thereof. Among the various compression techniques for compressing data, hybrid coding techniques combining probabilistic coding and temporal and spatial compression are known to be the most efficient, and these techniques have been proposed by the World Standards Organization for example. It is widely disclosed in the recommendations of MPEG-1 and MPEG-2.

Most hybrid coding techniques use motion compensated DCPM (Differential Pulse Code Modulation), two-dimensional Discrete Cosine Transform (DCT), quantization of DCT coefficients, VLC (variable length coding), and the like. The motion compensation DPCM determines the movement of the object between the current frame and the previous frame, and predicts the current frame according to the movement of the object to produce a differential signal representing the difference between the current frame and the predicted value. This can be done for example by Staffan Ericsson's Fixed and Adaptive Predictors for Hybrid Predictive / Transform Coding, IEEE Transactions on Communication, COM-33, NO.12 (1985, December), or A motion Compensated Interframe Coding by Ninomiy and Ohtsuka. Scheme for Television Pictures, IEEE Transactions on Communication, COM-30, NO.1 (January, 1982).

In general, two-dimensional DCT converts a digital image data block, for example, an 8x8 block, into a DCT conversion coefficient by using or removing spatial redundancy between image data. This technique is described in Chen and Pratt's Scene Adaptive Coder, IEEE Transactions on Communication, C0M-32, NO.3 (1984, March). The DCT conversion coefficient may be processed through a quantizer, a zigzag scan, a VLC, or the like to effectively reduce (or compress) the amount of data to be transmitted.

In more detail, the moved compensation DPCM predicts the current frame from the previous frame according to the motion of the object estimated between the current frame and the previous frame. The estimated motion may be represented by a two-dimensional motion vector representing the displacement between the previous frame and the current frame.

Typically, there are several approaches to estimating the displacement of an object. These are generally classified into two types, one of which is a block-by-block motion estimation method using a block matching algorithm and the other is a pixel-by-pixel motion estimation method using a pixel circulation algorithm.

In the motion estimation method for estimating the displacement of an object as described above, the displacement is obtained for each pixel by using the pixel-based motion estimation method. This method has the advantage of being able to estimate pixel values more accurately and easily handle scale changes (e.g., zooming, a movement perpendicular to the image plane), while the motion vectors are calculated for each pixel. Since it is determined, it is impossible to send substantially all the motion vectors to the receiver as large amounts of motion vectors occur.

In addition, in block-by-block motion estimation, an optimal matching block having a minimum error value is determined by comparing blocks with a predetermined size of a current frame by one pixel in a search range of a previous frame and comparing them with corresponding blocks. From this, the interframe displacement vector (the degree of block movement between frames) for the entire block is estimated for the current frame being transmitted. Here, for determining the similarity between two corresponding blocks between the current frame and the previous frame, the average absolute difference, the mean square difference, etc. are mainly used, as is well known in the art.

On the other hand, the image bit stream encoded by the encoding scheme as described above, that is, the encoding scheme such as motion compensation DCPM, two-dimensional DCT, DCT coefficient quantization and VLC (or entropy encoding) is transmitted to the output side of the image encoding system. The next transmission point stored in the buffer is sent to the transmitter for transmission to the remote destination. At this time, the transmission time point here is related to the size (ie, capacity) and transmission rate of the transmission buffer, and is controlled so that a malfunction (data overflow or data underflow) does not occur in the transmission buffer.

More specifically, the amount of bits generated for each frame at the time of encoding is changed due to various factors (e.g., the complexity of the image). In view of this, in the image encoding system, the average data rate can be kept constant. Control output buffer. That is, the video encoding system adjusts the amount of bits to be allocated in the current frame, based on the data fullness state information of the output side transmission buffer, to investigate the amount of bits generated before the frame currently encoded. In other words, in the conventional typical video encoding system, the amount of bits generated in the encoding system is adjusted by controlling the quantization step size (QP) substantially based on the data fullness state information of the output transmission buffer, that is, if the amount of bits generated before has been large, The bit generation amount is controlled by reducing the bit generation amount by adjusting the step size largely, and in the opposite case, by adjusting the quantization step size small to increase the bit generation amount.

However, as described above, the conventional method of adjusting the bit generation amount by adjusting the quantization step size based on the data fullness state information of the output side transmission buffer is performed when encoding and transmitting image data corresponding to each frame at the same data rate. In the case where the image to be encoded is complex (a large amount of high frequency components are generated), a large amount of bits is generated, which causes a problem that the quantization step size becomes large, resulting in severe image quality degradation in the reproduced image. The high frequency component generated here is a component that is substantially insensitive to human visual characteristics (a component that hardly affects the image quality of a reproduced video).

Accordingly, the present invention is to solve the above-mentioned problems of the prior art, in the coding system including MC-DCT, quantization, based on the quantization error value between the current frame and the reconstructed previous frame input for encoding A bit generation amount adjusting function that calculates the complexity of the image to be encoded and selectively adjusts the bit generation amount after encoding by selectively removing high frequency components of the input video signal using a one-dimensional low pass filter according to the calculation result. An object of the present invention is to provide an image encoding system.

In order to achieve the above object, the present invention provides a discrete cosine transform, quantization, and an error signal between an input current frame and a prediction frame obtained through motion estimation and compensation in units of macroblocks using the current frame and the reconstructed previous frame. Compression encoding is performed through encoding means including entropy encoding to generate an encoded bit stream, and the quantization has a bit generation amount adjusting function of adjusting a step size based on the fullness information of the bit stream stored in an output buffer. In an image encoding system, a quantization error value is calculated for each frame based on a pixel value difference signal between the current frame and a reconstructed previous frame, and the calculated quantization error values of the corresponding frame are averaged for a plurality of preset frames. Calculate the mean quantization error Quantization error calculating means; The calculated average error value of the plurality of frames is referred to as the complexity of the frame to be currently encoded by the encoding means, and the frequency pass of the current frame input for encoding based on the detected average error value of each frame is passed. Control means for generating a filter control signal for limiting the bandwidth; And based on the generated filter control signal, the input current frame is provided to the encoding means as the current frame signal for motion estimation and compensation, or one-dimensional low pass filtering is applied to the input current frame to pass the band. And a one-dimensional low pass filtering means for providing a frame signal from which a high frequency component is removed to the encoding means as a current frame signal for motion estimation and compensation by limiting the signal generation. To provide.

The above and other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a video encoding system having a bit generation amount adjusting function according to a preferred embodiment of the present invention.

As shown in the figure, the image encoding system of the present invention includes a first frame memory 100, a subtractor 110, an image encoding block 120, an entropy encoding block 130, a transmission buffer 140, and image decoding. Block 150, adder 160, second frame memory 170, current frame prediction block 180, quantization error calculation block 210, filter control block 220, and low pass filtering block 230. do.

Referring to FIG. 1, the input current frame signal is stored in the first frame memory 100 and then input to the low pass filtering block 230. In the low pass filtering block 230, the filter control block 220 described later is described. Selectively filter the frequency of the input frame signal according to a control signal (a logic signal of level 0 or 1) calculated based on the complexity of the post-encoding video provided from the input video signal, that is, the one-dimensional low pass filtering. Filtering the high frequency component of the (reduced portion of the comparative human eye), the detailed operation of this low-pass filtering block 230 will be described later in detail with reference to FIGS. Will be. Then, the current frame signal in which the high frequency component is not removed or removed from the input image is transmitted to the subtractor 110, the current frame prediction block 180, and the quantization error calculation block 210, which is a feature of the present invention, through the line L11. Each is provided.

First, the subtractor 110 performs motion compensation on a moving object provided from the current frame prediction block 180 described later through the line L19 from the current frame signal provided from the first frame memory 100 through the line L11. The predicted current frame signal is subtracted, and the resulting data, i.e., the error signal representing the differential pixel value, is obtained by means of discrete cosine transform (DCT) and one of the quantization methods well known in the art through the image coding block 120. By using, it is encoded into a series of quantized DCT transform coefficients. In this case, the quantization of the error signal in the image encoding block 120 is performed based on the quantization parameter QP determined according to the data fullness state information provided from the output transmission buffer 140 described later through the line L21. The size is adjusted.

Further, according to the present invention, the error signal on the line L11 is provided to the quantization error calculation block 210 to be described later, the quantization error calculation block 210 is described later through the input image signal for the current frame on the line L11 and the line L16. The quantization error value is calculated by using the image signal of the reconstructed previous frame provided from the second frame memory 170, that is, as a difference signal between the input image before encoding (DCT, quantization) and the reconstructed previous image after encoding. A quantization error value that may appear is calculated, and the calculated quantization error value is referred to as the complexity of the current frame to be encoded. In the present invention, the low-pass filtering block 230 adaptively (or selectively) removes high-frequency components relatively insensitive to human visual characteristics by using one-dimensional low-pass filtering based on the calculated complexity. do.

Meanwhile, although the detailed illustration in FIG. 1 is omitted, in the low pass filtering block 230, a control signal (0 or 1) based on the calculated quantization error value information provided from the filter control block 220 described later according to the present invention. 1D low pass filtering is used to determine the high frequency components that are relatively insensitive to human visual characteristics (ie, filtering). Accordingly, the present invention can appropriately adjust the quantization step size in the quantization step in the image encoding block 120 that causes deterioration of image quality in the reproduced video even in a complex image accompanied by an increase in the encoded bit generation amount. A detailed process of selectively limiting the passband of the input frame using the control signal generated based on the calculated quantization error value information will be described later in detail.

Next, the quantized DCT transform coefficients on line L13 are sent to entropy coding block 130 and image decoding block 150, respectively. Here, the quantized DCT transform coefficients provided to the entropy coding block 130 are encoded, for example, through a variable length coding scheme, and provided to the transmission buffer 140 on the output side. It is delivered to the transmitter not shown for transmission.

Meanwhile, the quantized DCT transform coefficients on the line L13 provided from the image coding block 120 to the image decoding block 150 are converted into reconstructed frame signals through inverse quantization and inverse discrete cosine transform, and then adder 160. The adder 160 adds the reconstructed frame signal from the image decoding block 150 and the predicted current frame signal provided from the current frame prediction block 180 described later through the line L19 to reconstruct. Generated previous frame signal, and the reconstructed previous frame signal is stored in the second frame memory 170. Accordingly, the immediately previous frame signal for every frame encoded through this path is continuously updated, and the reconstructed previous frame signal thus updated is provided to the current frame prediction block 180 described later for motion estimation and compensation. .

In addition, the reconstructed reconstructed previous frame signal stored in the second frame memory 170 is provided to the quantization error calculation block 210 described below through line L16 for calculating the complexity of the frame according to the present invention.

On the other hand, in the current frame prediction block 180, a signal of the current frame in which the high frequency component on the line L11 provided from the low pass filtering block 230 according to the present invention is selectively removed or the high frequency component is not removed is selected. A preset search range (eg, 16 × 16 or 32 × 32 search) of the previous frame reconstructed using the block matching algorithm based on the reconstructed previous frame signal on line L15 provided from one second frame memory 170 Range and predict the current frame in units of predetermined blocks (e.g., 8x8 or 16x16 DCT blocks) and then generate the predicted current frame signal on line L19 to add the subtractor 110 and the adder ( 160 respectively. At this time, the switch SW on the line L19 is turned on / off according to the control signal CS from the system controller (not shown). When the switch SW is turned on, the current encoding mode is the inter mode. In contrast, when off, the current encoding mode is intra mode. Accordingly, the subtractor 110 provides an error signal between the current frame signal and the predicted frame signal to the image encoding block 120 during inter-mode encoding, and transmits the current frame signal itself to the image encoding block 120 during intra-mode encoding. to provide.

In addition, the current frame prediction block 180 generates a set of motion vectors for each of the selected blocks (8 × 8 or 16 × 16 blocks) on the line L17 and provides the above-described entropy coding block 130. Here, the sets of motion vectors detected are predicted in the block of the current frame (8x8 or 16x16 block) and the preset search area (e.g., 16x16 or 32x32 search range) in the previous frame. It is the displacement between the most similar blocks. Accordingly, in the entropy coding block 130 described above, the quantized DCT transform coefficients on the line L13 together with the sets of the motion vectors on the line L17 generate a coded bit stream by encoding, for example, through a variable length coding technique. .

Meanwhile, the quantization error calculation block 210 of the present invention calculates the quantization error between the current frame on the line L11 input for encoding and the previous frame on the line L16 reconstructed and reconstructed after encoding, that is, input before encoding (DCT, quantization). A quantization error value is calculated for each frame that may appear as a difference signal between an image and a previous image reconstructed after encoding, and the calculated quantization error value is referred to as the complexity of the next frame to be encoded.

For example, assuming that one frame has a size of M × N, Ip is a picture of a previous frame reconstructed and reconstructed after protection, and Ic is a picture of a current frame input for encoding. Each pixel value at position (x, y) corresponds to Ip (x, y) and Ic (x, y), and the quantization error QE (Quantization Error) at this time is calculated by Equation (1) below. .

Here, the quantization error QE value calculated using Equation (1) may be a value that substantially reflects the complexity of the image data. Accordingly, the present invention uses the quantization error QE value calculated using the current frame and the previous frame reconstructed after encoding and used as the complexity of the next video signal input for encoding. In this case, the calculated quantization error QE value may be considered to be related to the information amount (bit generation amount) of the image. If the current encoded image is complex, the calculated quantization error QE value will be relatively large, and vice versa. The calculated quantization error QE value will be small.

In light of this point of view, the quantization error QE value between the current frame and the reconstructed previous frame calculated at the time of encoding may be interpreted as a value related to the amount of information to be transmitted. In calculating the average error AE value of each frame according to the present invention, since the encoding system stores the reconstructed previous frame signal required in the process of performing the motion estimation and compensation, such an encoding system. Using a part of, it is easy to implement the desired quantization error QE value which can be represented as the difference signal between the current frame and the reconstructed previous frame described above without any further calculation. Then, the quantization error QE value calculated by the quantization error calculation block 210 as described above is provided to the next filter control block 220.

On the other hand, the filter control block 220 generates a filter control signal for the one-dimensional filtering of the input image on the line L23 based on the average error QE value provided from the filter control block 210 to generate a low pass filtering block ( 230, that is, in the filter control block 220, when the calculated average error QE value input from the quantization error calculation block 210 is greater than a predetermined threshold value, a high level logic signal is generated to generate a low pass filtering block ( 230, and when the calculated average error AE value is smaller than the preset threshold, a low level logic signal is generated and provided to the low pass filtering block 230. In this case, the preset threshold may be set to an average value of the QE values generated during a predetermined time in the course of the encoder. As an example, when the frame rate of the video signal is 30, 30 QE values calculated for 1 second may be averaged and compared with the QE value generated every frame. That is, when the average value of the QE value generated during the previous 30 frames is C, in the filter control block 220, if the QE value generated in the current frame is larger than the average value C, the high level logic as the filter control signal on the line L23. A signal is generated, and when small, a low level logic signal is generated as the filter control signal on the line L23. Of course, the QE average value C for the previous 30 frames will be continuously updated by the QE value for every frame that is generated, at which time the oldest QE value of the QE values for the previous 30 frames is discarded.

Meanwhile, the low pass filtering block 230 passes the input video signal as it is, based on the filter control signal C from the filter control block 220 (that is, the subtractor 110 and the current frame through the line L11). Or a 1D low pass filtering to the input video signal to limit the passband, thereby generating a video signal on the line L11 from which high frequency components are relatively insensitive to human visual characteristics. do. That is, the low pass filtering block 230 removes high frequency components from the input image through one-dimensional low pass filtering when the filter control signal on the line L23 is a high level logic signal, and the filter control signal on the line L23 is low level logic. In the case of a signal, a high frequency component is removed from the input image without being removed.

More specifically, the low pass filtering method of an input video signal according to the present invention is performed by multiplying a low pass filter by an input video signal. That is, when the value of each pixel for an image of one block of N × M (for example, 8 × 8 blocks) is f (x, y), one block is shown as shown in FIG. 2 as an example. In the case of an 8x8 block, the position values x and y of the pixels in the horizontal and vertical directions have integer values between 0 and 7, and each value has a level value between 0 and 255. That is, as can be seen from the second diagram, the horizontal and vertical position values of each pixel of the 8x8 block have a value of f (0,0) to f (7,7).

On the other hand, as the low pass filter in the present invention, as shown in FIG. 3 as an example, the one-dimensional low pass filter coefficient is assumed to have seven orders. The low pass filter is an example of a low pass filter that band-limits the signal to have a frequency bandwidth of fs / 4 since the frequency bandwidth is fs / 2 when the sampling frequency of the input video signal is fs.

Accordingly, in the low pass filtering block 230, the low pass filtering of the video signal according to the present invention is performed by performing one-dimensional low pass filtering on each direction since the input video signal is a one-dimensional signal in the horizontal and vertical directions. Can be implemented. This process is illustrated in detail in FIG. 4, which shows the horizontal filtering at (0,4) and the vertical filtering at (3,0). In other words, if the low-pass filtered signal in the vertical direction at the position of (x, y) is z (x, y), it is calculated by the following expression (2).

In the above formula (2), T means the order of the filter, so T = 7. Thus, the u, v values have integer values between -3 and 3. In the above formula (2), the k (u) value is a filter coefficient value, and the f (x, y) value is a pixel value. If the (yu) value of f (x, yu) is less than 0 in Equation (2) above, the value is 0, or when the value is larger than the M-1 value corresponding to the image size of one frame, the M-1 value Do it. This is a method of setting the end value when the pixel value exists only in the area, that is, 0 to M-1, and this filtering method is a technique already known in the art.

Therefore, if filtering is performed at all pixel positions as in Equation (2) above, as shown in FIG. 4, for example, one-dimensional filtering in the horizontal and vertical directions can be obtained. In this case, according to the present invention, in the process of performing such filtering at all pixel positions, horizontal filtering may be performed first, and vertical filtering may be performed again or the order of the horizontal filtering may be performed again.

As a result, in the low pass filtering block 230, when the filter control signal on the line L23 is high level, that is, when it is determined that the image complexity is large, the low pass filtering block 230 applies one-dimensional low pass filtering to the input video signal to apply the pass band. By limiting, a video signal from which high frequency components are relatively insensitive to human visual characteristics is generated on the line L11, so that the subtractor 110 of FIG. 1, the current frame prediction block 180, and the quantization error calculation block 210 of the present invention. Will provide each.

Accordingly, in the image encoding block 120 of FIG. 1, in the case of a complex image, encoding (quantization) is performed in a state in which a high frequency component of an image that is relatively insensitive to human vision is removed through one-dimensional low pass filtering as described above. Therefore, the low frequency signal, which is a visually important component, can be encoded with less quantization error. If the passband of the frequency is not limited (high frequency component removal) as in the present invention despite the complicated image, as a result, the amount of bits generated after encoding increases and the quantization step size becomes large. In the low frequency band), a lot of quantization errors are generated, and as a result, image quality deterioration due to quantization will be caused in the playback image of the receiver.

As described above, according to the present invention, the current frame is to be encoded using the current frame input for encoding and the quantization error value information calculated for each frame using the previous frame reconstructed and reconstructed after encoding for motion estimation and compensation. If the current input image is a complex image based on the result of the calculation, and the current input image is a complex image, a high frequency component of the image insensitive to the human vision is firstly given by corresponding weight, and then the MC-DCT, quantization, etc. encoding is performed. By performing, the amount of bits generated after encoding can be effectively adjusted without excessive step size increase in the quantization step. Therefore, according to the present invention, it is possible to effectively reduce image quality deterioration due to quantization error inevitably present in a reproduced image when reconstructing and displaying an encoded image.

Claims (5)

Compression encoding through encoding means including discrete cosine transform, quantization, and entropy encoding for an error signal between an input current frame and a prediction frame obtained through macroblock motion estimation and compensation using the current frame and the reconstructed previous frame. And generating a coded bit stream, wherein the quantization is a bit generation amount adjusting function of adjusting a step size based on the full state information of the bit stream stored in an output buffer. Quantization error calculation means for calculating a quantization error value for each frame based on the pixel value difference signal between previous frames, and calculating the average quantization error value for a plurality of predetermined frames by averaging the calculated quantization error values of the corresponding frame. ; The calculated average error value of the plurality of frames is referred to as the complexity of the frame to be currently encoded by the encoding means, and the frequency pass of the current frame input for encoding based on the detected average error value of each frame is passed. Control means for generating a filter control signal for limiting the bandwidth; And based on the generated filter control signal, provide the input current frame to the encoding means as the current frame signal for motion estimation and compensation, or apply one-dimensional low-pass filtering to the input current frame to pass the band. And a one-dimensional low pass filtering means for providing a frame signal from which a high frequency component is removed to the encoding means as a current frame signal for motion estimation and compensation by limiting the signal generation. . The filter control signal of claim 1, wherein the filter control signal comprises: an average error value for a plurality of frames transmitted per second obtained by averaging the calculated average quantization error value of each frame, the average quantization error value of each frame, and the average error value Image coding system having a bit generation amount adjustment function, which is determined according to a standard deviation of a value. The image encoding system of claim 2, wherein the filter control signal is a logic signal of a high or low level. The image encoding system of claim 1, wherein the one-dimensional low pass filtering of the current frame is performed in units of 8 × 8 blocks. The method of claim 1 or 4, wherein the one-dimensional low pass filtering for the 8x8 block is performed sequentially in the horizontal-vertical or vertical-horizontal direction of each pixel position in the 8x8 block. An image encoding system having a bit generation amount adjustment function.