JP2001238215A - Moving picture coding apparatus and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a code quantity controller that is suitable for fixed bit rate and variable bit rate coding especially in real time with respect to high efficiency coding for a moving picture. SOLUTION: The moving picture coding apparatus that codes a moving picture is provided with a means 23 that detects a generated code amount of each picture, a means 22 that detects an average quantization scale of each picture, a measn 25 that detects a coding generating characteristic of at least a received picture in the received picture and a motion compensation prediction picture, a means 28 that detects a scene change from the coded picture characteristic, a means 24 that calculates image complexity on the basis of the generated code amount, the average quantization scale, the coded picture characteristic and the scene change information, and a means 14 that decides an assigned code amount of a picture to be succeedingly coded on the basis of the generated code amount, the image complexity, and the coded image characteristic and decides a quantization scale of a picture to be coded next from the assigned code amount.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to high-efficiency coding of moving pictures, and more particularly to a code amount control apparatus and method suitable for performing fixed bit rate and variable bit rate coding in almost real time.

[0002]

2. Description of the Related Art MPEG2 has already been defined as an international standard for technology for efficiently encoding moving images such as TV signals.
MPEG2 divides a “frame” image that constitutes a moving image into 16 × 16 pixel blocks called “macroblocks”. Each macroblock unit includes a reference image separated by a predetermined number of frames before or after a predetermined number of frames. A motion amount called a "motion vector" is obtained between the encoded images, and a "motion compensation prediction" technique for constructing the encoded image from the reference image based on the motion amount, and an error signal or code of the motion compensated prediction. "Transform coding" that compresses the amount of information on the coded image itself using DCT (Discrete Cosine Transform), a type of orthogonal transform
The technology is defined based on two image coding element technologies.

FIG. 7 shows an example of the configuration of a conventional MPEG-2 moving picture coding apparatus. FIG. 6 shows an example of an encoded picture structure. In the motion-compensated prediction, as in the coded picture structure shown in FIG.
It is composed of a combination of three types of pictures with different prediction methods, called pictures (forward prediction coding) and B pictures (bidirectional prediction coding).

As shown in FIG. 7, in transform coding,
The DCT unit 72 applies DCT to the coded image itself in the I picture, and to the output of the subtractor 71 which is the error signal of the motion compensation prediction by the motion compensator 79 in the P and B pictures. After the DCT coefficients obtained by the DCT unit 72 are quantized by the quantizer 73 under the control of the output of the code amount control unit 90, variable-length coding is performed together with other accompanying information such as motion vectors. Is the variable length encoder 75
The code string is output after being stored in the buffer 76 as a “bit stream”. At this time, the quantization scale is controlled by the code amount control unit 90 in accordance with the sufficiency of the buffer 76. On the other hand, the output coefficients of the quantizer 73 are supplied to an inverse quantizer 77 and an IDCT unit 78, and after being locally decoded, stored in a frame memory 81 for each block via an adder 80.

[0005] Since MPEG2 is variable-length coding, the amount of generated code (bit rate) per unit time is not constant. Therefore, it is possible to control to a required bit rate by appropriately changing the quantization scale at the time of quantization in the quantizer 73 for each macroblock. MPEG2
Test Model 5 proposes a fixed bit rate control method for making the generated code amount constant in GOP units. This fixed bit rate control method is a method suitable for an application requiring a constant transfer rate.

An outline of a fixed bit rate control method corresponding to the operation of the code amount control unit in FIG. 7 in Test Model 5 is as follows. Assuming that the target bit rate is BitRate, the number of frames per second is PictureRate, and the number of frames of one GOP (usually the interval between I pictures) as one encoding unit is N, the code amount R allocated to one GOP is represented by the following equation ( Given in 1). R = (BitRate / PictureRate) N (1)

The code amount R in the equation (1) is allocated to each image in the GOP. Here, for the image immediately after coding of each picture type, the generated code amount of one frame and the average quantization scale are calculated. The product is obtained as screen complexity (Complexity) Xi (I picture), Xp (P picture), Xb (B picture), and the image in the GOP including the image to be encoded is uniformly the screen complexity (Complexity) Is determined, the target allocated code amount of the image to be encoded is determined.

[0008] P, whose encoding is not completed in the current GOP,
It is assumed that the number of frames of the B picture is Np and Nb, and the setting ratio of the quantization scale of the P picture and the quantization scale of the B picture is Kp and Kb. At this time, the target allocation code amounts Ti, Tp, and Tb of the I, P, and B picture types are given by the following equations (2), (3), and (4). MAX [A, B] indicates an operation of selecting the larger one of A and B.

(I picture) Ti = MAX [A, B] A = R / (1 + Np.Xp / (Xi.Kp) + Nb.Xb / (Xi.Kb)) B = BitRate / (8.PictureRate) ) (2) (P picture) Tp = MAX [C, D] C = R / (Np + Nb · Kp · Xb / (Xp · Kb)) D = BitRate / (8 · PictureRate) (3) (B picture ) Tb = MAX [E, F] E = R / (Nb + Np ・ Kb ・ Xp / (Xb ・ Kp)) F = BitRate / (8 ・ PictureRate) (4)

Note that the value of the code amount R is calculated by subtracting the generated code amount of the frame every time one frame coding is completed.
At the beginning of P (I picture), the value of equation (1) is added.
Next, the target allocation code amount determined by the above equations (2), (3), (4),
The quantization scale of each macroblock is determined based on the generated code amount of each macroblock detected by the buffer 76.

On the other hand, there is a variable bit rate control method as a method suitable for applications in which a variable transfer rate is possible such as DVD-Video. JP-A-6-141298 discloses an encoding device using variable bit rate control.
In this device, first, provisional encoding is performed on an input moving image using a fixed quantization scale, and the generated code amount is counted for each unit time. Next, the target transfer rate of each part is set based on the generated code amount at the time of provisional encoding so that the generated code amount of the entire input moving image becomes a required value. Then, while performing control so as to match the target transfer rate, the second encoding, that is, actual encoding, is performed on the input moving image.

However, in the above-mentioned conventional example, encoding must be performed at least twice in order to obtain an output bit stream. In applications requiring real-time performance, a two-pass system such as this apparatus is used. Variable bit rate control cannot be used.

On the other hand, there is also a variable bit rate control method for encoding a moving picture almost in real time, that is, a variable bit rate control method of a one-pass system. Japanese Patent Application Laid-Open No. H10-164577 discloses an encoding apparatus using a one-pass variable bit rate control method, for example, in FIG. FIG. 8 shows an example of the configuration of a moving picture coding apparatus in this conventional example. The same components as those in FIG. 7 are denoted by the same reference numerals, and description thereof will be omitted.

In this conventional apparatus, the code amount stored in the buffer 76 is supplied to a generated code amount detector 83, and the generated code amount by the generated code amount detector 83 and the quantizer 73
Quantized scale from the average quantized scale detector 82
And the product of the average quantization scale in the screen by the average quantization scale detector 82 and the average value of the quantization scale in the screen is calculated as “screen complexity” by the screen complexity detector 84, and the average value of the past screen complexity is calculated. Based on the current screen complexity ratio to
The variable bit rate control is realized by the code amount controller 74 by determining the target generated code amount or the target quantization scale of the screen.

[0015]

In the above conventional example, the coding control is performed on the assumption that the screen complexity of an image to be coded is about the same as the screen complexity of the same picture type coded immediately before. Is being done. However, when a large change such as a scene change occurs in the input moving image, not only does the screen complexity itself change because the nature of the image changes before and after the change point, but also in P and B pictures for which predictive coding is performed. In addition, since almost no prediction is made between the image before and after the change occurs, almost all macroblocks in the image immediately after the change often become intras for performing intra-frame coding.

In such an image, compared to the image before the change,
In the above-described conventional example, the code amount is allocated based on the screen complexity of the image before the change, despite the fact that the actual screen complexity becomes extremely high. There is a problem that the image scale is increased, and as a result, the image quality is deteriorated immediately after the change point.

On the other hand, in an image to be predicted next from the image immediately after the change, the prediction becomes relatively easy except when the image changes little by little, and the actual screen complexity is lower than that of the image immediately after the change. Nevertheless, in the above conventional example, the code amount is allocated based on the high screen complexity immediately after the change, so that the allocated code amount becomes excessive and the quantization scale drops unnecessarily, and the code amount is reduced. However, there is a problem that waste occurs. The present invention solves the above-described problems, and enables a fixed bit rate and a variable bit rate of a one-pass method to perform more appropriate code amount allocation even when a large change such as a scene change occurs in an input moving image. It is intended to realize a control method.

[0018]

According to a first aspect of the present invention, there is provided an image processing apparatus comprising a motion compensation predicting means, an orthogonal transform means, a quantizing means, and a variable length coding means. A moving image encoding apparatus for detecting and encoding a generated code amount of each image of the input moving image; a unit for detecting an average quantization scale of each image of the input moving image; Means for detecting at least an encoded image characteristic of the input moving image out of an image and a motion-compensated predicted image generated by the motion-compensated predicted image, and an encoded image characteristic detected by the means for detecting the encoded image characteristic Means for detecting scene change image information from
From the generated code amount detected by the means for detecting the generated code amount, from the average quantization scale detected by the means for detecting the average quantization scale, the actual screen complexity of a past image is calculated, and the calculated From the actual screen complexity of the past image, the coded image characteristic detected by the coded image characteristic detecting means, and the scene change image information detected by the scene change image information detecting means, the current image is obtained. Means for calculating the estimated screen complexity of the image, and means for calculating the estimated screen complexity of the past image calculated by the means for calculating the actually measured screen complexity, or the means for calculating the estimated screen complexity of the current image. The estimated screen complexity of the current image, the generated code amount detected by the means for detecting the generated code amount, and the average quantization scale. From the average quantization scale detected by the means for detecting the coded image characteristic and the coded image characteristic detected by the means for detecting the coded image characteristic. Means for determining a quantization scale of the image to be coded next, and a moving picture coding apparatus according to claim 2, wherein In the encoding apparatus, the means for calculating the estimated screen complexity of the current image includes: a coded image characteristic of the detected current image; and a picture type (I picture, P picture
Picture, B-picture) by multiplying the actually measured screen complexity in the immediately preceding image by a predetermined function having a ratio to the coded image characteristic detected in the image immediately before the previous image. The present invention provides a moving picture coding apparatus for calculating the scene change picture information in which the current picture is the detected scene change picture information in the moving picture coding apparatus according to claim 1. When it is determined that the current image is a scene change image or a next image after the scene change image, a predetermined function having a factor of the ratio with the encoded image characteristic is changed, and the current image is changed to a next image of the scene change image. If the image is determined to be an image, the immediately preceding image of the same picture type (I picture, P picture, B picture) as the current image is replaced with the same picture type as the current image. A moving image encoding apparatus characterized in that the moving image is changed to an image before a scene change image of the moving image. The invention according to claim 4 is the moving image encoding apparatus according to claim 1, wherein The screen complexity used in the means for determining the assigned code amount of the image to be encoded is determined at least as the current image is a scene change image or the next image of the scene change image according to the detected scene change image information. In this case, the present invention provides a moving picture coding apparatus characterized by using the estimated screen complexity of the current picture, and the invention of claim 5 is described in any one of claims 1 to 4. Means for determining an assigned code amount of an image to be encoded next from the generated code amount and the actually measured screen complexity or the estimated screen complexity. The ratio of the estimated picture complexity to the average value in a certain period of time,
Providing a moving image encoding apparatus characterized by determining the allocated code amount by multiplying the average allocated code amount,
According to a sixth aspect of the present invention, in the moving image encoding apparatus according to any one of the first to fifth aspects, the means for detecting an encoded image characteristic of the input moving image or the motion-compensated predicted image includes the input image characteristic. A picture type (I-picture, P-picture, and P-picture), comprising: means for detecting image characteristics of a moving image; means for detecting image characteristics of an error image of a motion-compensated prediction image; and means for detecting a motion vector characteristic in motion-compensated prediction. (B picture) and a constant for each prediction mode are multiplied by the respective characteristic values of the detected image characteristics of the input moving image, the image characteristics of the error image of the motion compensated prediction image, and the motion vector characteristics in the motion compensation prediction. Means for determining the coded image characteristic and detecting the scene change image information, the current image and the same picture Image characteristics of the input moving image detected by the means for detecting the coded image characteristics and image characteristics of an error image of the motion-compensated predicted image between images immediately before the current type (I picture, P picture, B picture). ,
And calculate the ratio of the respective characteristic values of the motion vector characteristics in the motion compensation prediction, if the ratio of the characteristic values exceeds a certain range determined for each characteristic value, the current image and the scene change image Determined and the same picture type as the scene change image (I picture, P picture, B picture)
A moving picture coding apparatus characterized in that a next picture of a moving picture is determined as a next picture of a scene change picture. And a moving image encoding method for encoding having a variable length encoding step, wherein a step of detecting a generated code amount of each image of the input moving image,
Detecting an average quantization scale of each image of the input moving image; and detecting at least an encoded image characteristic of the input moving image among the input moving image and the motion-compensated predicted image generated by the motion-compensated prediction step. , Detecting scene change image information from the coded image characteristics detected by the step of detecting the coded image characteristics, generating code amounts detected by detecting the generated code amounts, and averaging the generated code amounts. From the average quantization scale detected by the step of detecting the quantization scale, the actually measured screen complexity of the past image is calculated, and the actually measured screen complexity of the calculated past image and the encoded image characteristic are detected. Detecting the encoded image characteristics detected in the step and the scene change image information. Calculating the estimated screen complexity of the current image from the scene change image information detected by the method, and the measured screen complexity of the past image calculated by the step of calculating the measured screen complexity, or the current image. The estimated screen complexity of the current image calculated by the step of calculating the estimated screen complexity, the generated code amount detected by the step of detecting the generated code amount, and the average quantization scale are detected by the step of detecting the average quantization scale. The average quantization scale obtained, and the allocated code amount of the image to be next encoded is determined from the encoded image characteristics detected by the step of detecting the encoded image characteristics, and the next encoding is performed based on the allocated code amount. Determining a quantization scale of an image to be encoded. In the moving picture coding method according to claim 7, the step of calculating the estimated screen complexity of the current image includes the step of calculating the coded image characteristic of the detected current image and the same picture type ( (I picture, P picture, B picture) The estimation is performed by multiplying the measured screen complexity in the immediately preceding image by a predetermined function having a ratio to the coded image characteristic detected in the image immediately before the previous image. A moving picture coding method comprising calculating screen complexity is provided. According to a ninth aspect of the present invention, in the moving picture coding method according to the seventh aspect, the current image is detected. If it is determined by the scene change image information that the image is a scene change image or an image next to the scene change image, a predetermined function having a ratio to the encoded image characteristic as a factor is changed, and If the current image is determined to be the next image of the scene change image, the image immediately before the same picture type (I picture, P picture, B picture) as the current image is replaced with the same picture type as the current image. A moving picture coding method characterized by changing to a picture before a scene change picture of the moving picture coding method according to the present invention. The screen complexity used in the step of determining the allocated code amount of the image to be converted is at least that the current image is determined to be a scene change image or an image next to the scene change image by the detected scene change image information. In such a case, the present invention provides a moving picture coding method characterized by using the estimated screen complexity of the current picture. 11. The moving image encoding method according to claim 10, wherein an allocated code amount of an image to be encoded next is determined from the generated code amount and the measured screen complexity or the estimated screen complexity. A moving image coding method, wherein a ratio of the estimated screen complexity to an average value of the measured screen complexity for a certain period is multiplied by an average allocated code amount to determine the allocated code amount. According to a twelfth aspect of the present invention, in the moving image encoding method according to any one of the seventh to eleventh aspects, the step of detecting an encoded image characteristic of the input moving image or the motion-compensated predicted image includes: Detecting an image characteristic of the input moving image, detecting an image characteristic of an error image of the motion-compensated prediction image, and detecting a motion vector characteristic in the motion-compensated prediction. It is composed of that step, picture type (I picture, P
Picture, B picture) and prediction mode.
The coded image is obtained by multiplying each of the detected image characteristics of the input moving image, the image characteristics of the error image of the motion-compensated prediction image, and the respective characteristic values of the motion vector characteristics in the motion-compensated prediction, and adding the product. Determining a characteristic and detecting the scene change image information may include determining the current image and the same picture type (I picture,
(P picture, B picture), the image characteristics of the input moving image detected by the step of detecting the encoded image characteristics, the image characteristics of the error image of the motion compensated prediction image, and the motion compensation prediction Calculate the ratio of the respective characteristic values of the motion vector characteristics in, when the ratio of the characteristic values exceeds a certain range determined for each characteristic value, determine the current image as a scene change image, It is an object of the present invention to provide a moving picture coding method characterized in that the next picture of the same picture type (I picture, P picture, B picture) as a scene change picture is determined as the next picture of the scene change picture.

Therefore, according to the present invention, in a moving picture coding apparatus provided with means for motion compensation prediction such as MPEG2, orthogonal transformation, quantization and variable length coding, the generated code amount of each picture and the average quantization In order to detect the scale and distribute the allocated code amount in a predetermined section to each image in between, a predetermined operation is performed on the product of the generated code amount of each image and the average quantization scale, and the already coded image is Obtain the actual measurement screen complexity. At the same time, the coded image characteristics (activity) of each image are detected, the ratio of the coded image characteristics to the immediately preceding image of the same picture type is calculated, and a scene change is detected from the coded image characteristics.

The coded image characteristics of each image include, in addition to the activity of the input image, an error image in motion compensation prediction or a difference image between an encoded image and a reference image in motion vector detection, and a motion vector for P and B pictures. And one or more of them are used in combination. If the value of each element of the encoded image characteristic exceeds a predetermined range, the image is determined to be a scene change image.

For an image determined to be a scene change image for each picture type, an estimated screen complexity is calculated from a predetermined function having the above-described ratio of the coded image characteristics as a factor and the actually measured screen complexity of the immediately preceding image. Since the estimated screen complexity is larger than the actually measured screen complexity, a larger code amount can be allocated to the scene change image.

Conversely, if the immediately preceding image is a scene change image, the actually measured screen complexity calculated when assigning the code amount of the immediately preceding image, that is, the complexity before the scene change, is used instead of using the actually measured screen complexity. Estimating the screen complexity of an image to be encoded from the image before the scene change and a predetermined function whose factor is the ratio of the encoded image characteristic of the image to be encoded to the image to be encoded. Accordingly, it is possible to prevent an unnecessarily large code amount from being allocated to the image next to the scene change image.

In the case of one-pass variable bit rate control, the estimated screen complexity of the image to be encoded is calculated from the ratio of the encoded image characteristics and the actually measured image complexity of the immediately preceding image. For an image determined to be a scene change image in each picture type, an estimated screen complexity is calculated from a predetermined function having the above-described ratio of encoded image characteristics as a factor and the actual measured screen complexity of the immediately preceding image. The degree is added more than usual.

Conversely, when the immediately preceding image is a scene change image, instead of the actually measured screen complexity, the actually measured screen complexity of the image before the scene change is calculated when assigning the code amount of the immediately preceding image, or By estimating the screen complexity of the encoded image using the estimated screen complexity before addition in the scene change image, the value of the screen complexity in the image following the scene change image can be optimized.

In the variable bit rate control of the one-pass method, the calculated estimated screen complexity and the ratio of the average screen complexity for each picture type in a certain section are reflected in the code amount allocation in a predetermined section based on the target bit rate. Accordingly, it is possible to allocate a code amount in a predetermined section corresponding to a scene change.

[0026]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) A first embodiment of a moving picture coding apparatus and method according to the present invention will be described.
This will be described below with reference to the drawings. The first embodiment of the moving picture coding apparatus of the present invention shown in FIG.
T unit 12, quantizer 13, code amount controller 14, variable length encoder 15, buffer 16, inverse quantizer 17, IDC
T unit 18, motion compensation predictor 19, adder 20, frame memory 21, average quantization scale detector 22, generated code amount detector 23, screen complexity calculator 24, screen complexity memory 24M, image characteristic detector 25, and a scene change detector 28.

The first embodiment is a case where the present invention is applied to constant bit rate coding. In the following description, i corresponds to an I picture, p corresponds to a P picture, and b corresponds to a B picture. It is assumed that the original image has been previously divided into macroblock units by an image block divider (not shown). With respect to the divided source image, the motion compensation prediction is not performed for the I picture, and the source image block itself is sent to the DCT unit 12 via the subtractor 11, and after the DCT, the quantizer 13 controls the code amount controller 14. Quantized by the quantization scale sent from.

The quantized signal is converted into a code by the variable-length encoder 15 and adjusted by the next buffer 16 to output the code. On the other hand, the output coefficient of the quantizer 13 is locally decoded by the inverse quantizer 17 and the IDCT unit 18, and the output of the motion compensation predictor 19 is added to the frame memory 21 for each block without being added by the adder 20. Can be stored.

As for the P and B pictures, the divided original moving picture and predetermined locally decoded picture blocks stored in the frame memory 21 are supplied to the motion compensation predictor 19, where the motion vector detection and motion compensation are performed. Then, the pixel difference between the predicted image block and the original image block is calculated by the subtractor 11, and the error image block as the difference value is supplied to the DCT unit 13.

Thereafter, the DCT unit 1 operates similarly to the I picture.
2, the difference value is subjected to DCT, and the quantizer 13 controls the code amount controller 1.
After being quantized by the quantization scale sent from 4, it is converted into a code by the variable-length encoder 15, adjusted by the next buffer 16, and output. On the other hand, the output coefficient of the quantizer 13 is the inverse quantizer 17 and the IDCT
After the local decoding in (1) and (2), the predicted image block is added for each pixel by the adder 20 and stored in the frame memory 21 for each block.

For each picture, the quantizer 13
Is sent to the average quantization scale detector 22, where the quantization scale for one frame is added, and the average quantization scale for one frame is calculated. On the other hand, the generated code amount is monitored in the buffer 16, and the value is supplied to the generated code amount detector 23. The generated code amount detector 23 adds the generated code amount for each frame, and detects the generated code amount for one frame.

The average quantization scale and generated code amount detected for each frame are respectively calculated by the screen complexity calculator 24.
Is supplied for each frame. The screen complexity calculator 24 performs a predetermined operation after multiplying the average quantization scale of each supplied frame by the generated code amount, and performs MPEG2Te
The measured screen complexity Xi-p, Xp-p, Xb-p of each frame corresponding to Complexity in st Model 5 is obtained.

On the other hand, the image characteristic detector 25 supplies a moving image obtained by dividing the original image at the time of input, detects an activity which is a parameter indicating image characteristics for each frame of the moving image in units of macroblocks, and for each frame. The result is added to the screen complexity calculator 24 and the scene change detector 28 for each frame.

Here, the operation of detecting the image characteristics by the image characteristic detector 25 is performed prior to the actual encoding operation. As parameters indicating the image characteristics, variance of luminance values, difference values between pixels, and the like can be considered, but other parameters may be used as long as they indicate image characteristics.

In the scene change detector 28, the activities ACTi, ACTp, AC of the current image to be encoded are
Tb and the ratio of the activities ACTi-p, ACTp-p, and ACTb-p of the previously encoded image of the same picture type (ACTi / AC
Ti-p), (ACTp / ACTp-p), and (ACTb / ACTb-p) are calculated.

When the calculated ratio (current activity / previous activity) exceeds a certain range, for example, (Ratio) <Amin or (Ratio)> Amax (0 <Ami
In the case of n <1, Amax> 1), the scene change detector 28 determines that a scene change has occurred between these two images, and sends the position information of this frame to the screen complexity calculator 24. When it is determined that a scene change has occurred, the current image to be encoded is hereinafter referred to as a scene change image.

When the position information of the frame indicating that the scene change has occurred is sent to the screen complexity calculator 24, the screen complexity calculator 24 compensates for the increase in the amount of codes generated in the scene change image for the P and B pictures. (Compensation) of the screen complexity is performed.

Here, the measured screen complexity previously obtained is stored in the screen complexity memory 24M for use in estimating the screen complexity of the next same picture type. The estimated screen complexity Xp, Xb in the scene change image is given by the following equation (5)
~ (8) Note that f1 to f4 are predetermined functions having factors of the activity ratio (ACTp / ACTp-p) and (ACTb / ACTb-p). As an example of the predetermined function,
f1 to f4 = (ACTk / ACTk-p) · Ck1 + Ck2 (where k = p or
b) is suitable, but not limiting.

(P picture) When (ACTp / ACTp-p)> Amax Estimated screen complexity Xp = Xp-p · f1 (5) When (ACTp / ACTp-p) <Amin Estimated screen complexity Xp = Xp -p · f2 (6) (B picture) When (ACTb / ACTb-p)> Amax Estimated screen complexity Xb = Xb-p · f3 (7) When (ACTb / ACTb-p) <Amin Estimated screen complexity Degree Xb ＝ Xb-p ・ f4 (8)

For example, Amax = 1.5, Amin = 0.67, f1 = (AC
Tp / ACTp-p) ・ 2, f2 = 1, f3 = (ACTb / ACTb-p) ・ 2, f4
By setting = 1, if the activity increases, a value twice the activity ratio is given as the estimated screen complexity, and if the activity decreases, the same screen complexity as before the scene change is estimated. P and B by scene change
An increase in the amount of generated code due to an increase in intra macroblocks can be compensated for in the picture.

For the P or B picture of the same picture type following the scene change image, the screen complexity is calculated based on the screen complexity Xp-p, Xb-p calculated from the increased generated code amount of the scene change image. When calculating the degree,
The value becomes larger than necessary.

Therefore, instead of Xp-p and Xb-p, actual screen complexity Xp-pold and Xb-pold (stored in screen complexity memory 24M) used for the scene change image, Activity ACTp of the image targeted for screen complexity
-pold, ACTb The screen complexity is optimized by estimating the screen complexity from Equations (9) and (10) from the activity ACTp, ACTb of the current image to be coded, and the current image activity ACTp, ACTb.

(P picture) Xp = Xp-pold · (ACTp / ACTp-pold) (9) (B picture) Xb = Xb-pold · (ACTb / ACTb-pold) (10) The screen complexity memory 24M Stores the activity of the image of the last several frames of each picture type, in addition to the actually measured screen complexity in the scene change image.

In the first embodiment, the scene change image of the P and B pictures, the estimated screen complexity for the image following the scene change image, and the actual measurement screen for the other normal P, B and I pictures MPEG2Te complexity
By substituting into Xi, Xp, and Xb of equations (2) to (4) of st Model 5 and determining the target allocation code amounts Ti, Tp, and Tb of the image to be coded, the code amount corresponding to the scene change is obtained. Assignments can be made. Based on the target allocated code amount and the generated code amount of each macroblock detected by the buffer 16, the quantization scale of each macroblock is determined using the method of MPEG2 Test Model 5.

The activity of each macroblock is also sent from the image characteristic detector 25 to the code amount controller 14, and the adaptive quantization control for changing the quantization scale of each macroblock based on the activity in the MPEG2 Test Model 5 is performed. Although used, this adaptive quantization control need not be performed.

The quantization scale of each macroblock output from the code amount controller 14 is sent to the quantizer 13, and the current image (a divided original image after DCT or an error image block for motion compensation prediction) is obtained. After being quantized by the quantizer 13 at this quantization scale, variable-length encoded by the variable-length encoder 15 and adjusted by the next buffer 16, the code is output. The quantization scale for each macroblock of the quantizer 13 and the generated code amount monitored by the buffer 16 are sent to the average quantization scale detector 22 and the generated code amount detector 23, respectively, and the code amount control of the next picture is performed. Used for

Although the present embodiment is an example in which the present invention is applied to the MPEG2 Test Model 5, the present invention is not limited to this.
It is applicable to other rate controls. For example, if the allocated code amount Sk of the next image to be encoded is determined from the generated code amount of a past image of the same picture type whose activity is ACTk-p (k = i, p, b), the activity is ACTk. Assuming that the image is a scene change image, the corrected allocated code amount Sk 'is expressed by the following equation (11). Sk ′ = Sk · fk (ACTk / ACTk-p) (11) where k = i, p, b fk is a function with (ACTk / ACTk-p) as a factor

(Second Embodiment) Next, a second embodiment of the moving picture coding apparatus according to the present invention will be described below with reference to FIG. The second embodiment is a case where the present invention is applied to one-pass variable bit rate encoding. Compared with the first embodiment, an average screen complexity calculator 29 is added, and the screen complexity calculation is performed. Since only the configuration and operation of the code amount controller 14 and the code amount controller 14 are different, the description of the other parts is omitted.

As in the first embodiment, the screen complexity calculator 24 calculates the average quantization scale and generated code amount of each frame, the position information of the frame at which a scene change has occurred, the image characteristics of each frame, and the like. That is, an activity is supplied. The screen complexity calculator 24 multiplies the average quantization scale of each supplied frame by the generated code amount and then performs a predetermined conversion on the multiplication result. Based on this, the measured screen complexity of each frame is calculated. Desired.

The measured screen complexity is sent to the average screen complexity calculator, where the values within a certain period are added for each coded picture type, and then divided by the number of frames of the same picture type within that period. , I, P, and B picture types, the average screen complexity Xi-ave (I picture), Xp-ave (P picture), and Xb-ave (B picture) are calculated.

In the above-mentioned fixed period, there may be a fixed number of frames, for example, 15 frames or 300 frames, which are predetermined in time before the picture for which encoding has just been completed. In some cases, such as from a frame to an image for which encoding has just finished, the number of frames is sequentially increased. Note that, even in the case of the former fixed number of frames, if the number of encoded frames does not satisfy the predetermined certain period, the number of frames is sequentially increased similarly to the latter.

Next, the screen complexity calculator 24 estimates the screen complexity from the measured screen complexity calculated, the average screen complexity calculated by the average screen complexity calculator 29, and the activity. Screen complexity of the current image to be encoded
Xi, Xp, Xb are the current image activities ACTi, AC
Tp, ACTb, screen complexity Xi-p, Xp-p, Xb-p of the previously coded image of the same picture type, activity ACTi-p, ACTp- of the previously coded image of the same picture type
From p and ACTb-p, it can be estimated by the following equations (12), (13), and (14).

(I picture) Screen complexity of current image Xi = Xi-p · (ACTi / ACTi-p) (12) (P picture) Screen complexity of current image Xp = Xp-p · (ACTp / (ACTp-p) (13) (B picture) Screen complexity of the current image Xb = Xb-p · (ACTb / ACTb-p) (14)

In the initial state, if there is no frame for which the same picture type has been coded, the screen complexity and activity of each picture type image are obtained in advance for some images, and the average is calculated. The average value may be statistically averaged in accordance with the frequency of occurrence of a moving image, and may be used as an initial value.

On the other hand, if the current image to be encoded is a scene change image, the screen complexity calculator 24 recalculates the estimated screen complexity for P and B pictures in order to allocate a larger amount of code. . Here, the actually measured screen complexity of the image of the same picture type encoded immediately before is stored in the screen complexity memory 24M for use in estimating the screen complexity of the next same picture type.

The estimated screen complexity Xp, Xb in the scene change image is obtained in the same manner as in the equations (5) to (8) of the first embodiment. Note that f1 to f4 are the activity ratios (ACTp / ACTp
-p) and (ACTb / ACTb-p) are the predetermined functions, but a function different from that of the first embodiment may be used.

Further, for the P or B picture of the same picture type following the scene change image, the screen complexity memory 24M is used in the same manner as in the equations (9) and (10) of the first embodiment.
Screen complexity Xp-pold and Xb-pold stored in the screen, and the activity AC of the image subject to the measured screen complexity
By estimating the screen complexity from Tp-pold, ACTb-pold, and the current image activity ACTp, ACTb to be encoded, the screen complexity is optimized.

In the second embodiment, the estimated screen complexity is calculated regardless of the presence or absence of a scene change.
Is not the measured screen complexity, but is changed to the estimated screen complexity Xp-old, Xb-old before recalculation obtained by the above formulas (13) and (14), and the same picture following the scene change image The following (15) (1)
It may be calculated by the formula 6). In this case, it is not necessary to switch the images stored in the screen complexity memory 24M depending on the presence or absence of a scene change, so that the apparatus can be simplified.

(P picture) Screen complexity of the current image Xp = Xp-old · (ACTp / ACTp-p) (15) (B picture) Screen complexity of the current image Xb = Xb-old · (ACTb / ACTb-p) (16)

The estimated screen complexity Xi, Xp, Xb of the current image to be encoded, calculated in this way, and the average screen complexity Xi-ave, Xp-ave, Xb-av of each picture type
e is sent to the code amount controller 14.

Next, the code amount controller 14 (from now on)
The setting (determination) of the allocated code amount of the image to be encoded and the setting (determination) of the quantization scale for variable bit rate control are performed. The target average bit rate is BitRate, the number of frames per second is PictureRate, and one encoding unit is 1 GOP
If the number of frames (usually the interval between I pictures) is N, 1G
The average allocated code amount Rave of the OP is given by the following equation (17).

Average allocated code amount of 1 GOP Rave = (BitRate / PictureRate) · N (17) If Rave in the above equation is the required allocated code amount of 1 GOP at the time of average screen complexity, the current image to be encoded is 1GOP including
Is assumed to be uniformly equal to the estimated screen complexity of the current image obtained by the screen complexity calculator 24, the required allocated code amount Rc of 1 GOP required to keep the image quality constant is (18) (19) (20)

(I picture) Required allocated code amount Rc = Rave · (Xi / Xi-ave) (18) (P picture) Required allocated code amount Rc = Rave · (Xp / Xp-ave) (19) (B picture ) Necessary allocated code amount Rc = Rave ・ (Xb / Xb-ave) (20)

By appropriately allocating the required allocated code amount Rc in the above equation to each picture of 1 GOP, the target code amount of the current image to be coded is calculated, and the quantization scale of each macroblock is determined. For example, MPEG2 Test Mod
Using the method of el5, the estimated screen complexity Xi, Xp, Xb of the current image obtained above is replaced by Xi, Xp, Xb in equations (2) to (4), and the allocated code amount Rc is assigned by (2) to Substituting into R in equation (4), the target allocated code amounts Ti, Tp, Tb of the image to be encoded are determined.

However, in the second embodiment, since it is not necessary to set a constant code amount for each GOP, the value of Rc is determined by the generated code amount of each frame as shown by R in equations (2) to (4). Subtract or go
There is no need to add at the beginning of P. Also, Np, N in equations (2) to (4)
b is always a constant value (the value at the beginning of the GOP).

Thereafter, as in the first embodiment, the quantization scale of each macro block is determined based on the target allocated code amount and the generated code amount of each macro block detected by the buffer 16, and Adaptive quantization control for changing the quantization scale in accordance with the activity of each macroblock according to. In this way, even in the variable bit rate control of the one-pass method, it is possible to allocate the code amount corresponding to the scene change.

Third Embodiment Next, a third embodiment of the moving picture coding apparatus according to the present invention will be described below with reference to FIG. The third embodiment is also a case where the present invention is applied to one-pass variable bit rate coding.
As compared with the embodiment, the configuration and operation of the image characteristic detector 25 and the scene change detector 28 shown in FIG.
Only the content of the activity signal sent from the image characteristic detector 25 to the scene change detector 28 and the content of the scene change information sent from the scene change detector 28 to the screen complexity calculator 24 are different. FIG. 3 is different from FIG. 2 in that a motion compensation signal is supplied from the motion compensation predictor 19 to the image characteristic detector 25, and description of other parts is omitted.

The image characteristic detector 25 shown in FIG.
Detector 25A, ACTpred detector 25B, ACTmv detector 25
C, accumulators 25D, 25E, and 25F, and a picture activity calculator 25G. The scene change detector 28 is a SCcur detector 28.
A, an SCpred detector 28B, an SCmv detector 28C, and a scene change determiner 28D.

In the embodiment shown in FIGS. 3 and 4, the input to the image characteristic detector 25 is divided into macroblock units as in the second embodiment, since no motion compensation prediction is performed for an I picture. Is input, only activity (ACTcur), which is a parameter indicating image characteristics in macroblock units, is detected and added in frame units.
It is sent to the screen complexity calculator 24 as an I-picture activity ACTi. The output value of the ACTcur detector 25A is added to each frame by the accumulator 25D, sent to the scene change detector 28, and used for scene change determination.

On the other hand, the input to the image characteristic detector 25 shown in FIG. 4 is, in the case of P and B pictures, an error image in macroblock-based motion compensation prediction or a motion vector detection in addition to the divided original moving image. The difference image between the coded image and the reference image and the motion vector used in the motion compensation prediction are input from the motion compensation predictor 19 shown in FIG. From the divided original image, the activity (actual image) ACTcur
Is detected.

On the other hand, an error image in motion compensation prediction in macroblock units or a difference image between a coded image in motion vector detection and a reference image is subjected to a sum of absolute values or a sum of squared errors in the error image.
Is detected as Further, the motion vector used in the motion compensation prediction is detected as a motion vector activity ACTmv, for example, by taking the absolute value of a difference between each component and an adjacent macroblock.

Then, the macroblock activity ACTmb is calculated for each macroblock by the operation of the following equation (21), and it is added for one frame to calculate the screen complexity as the P and B picture activities ACTp and ACTb. To the vessel 24. The values of the ACTcur, ACTpred, and ACTmv detectors 25A, 25B, and 25C are stored in accumulators 25D and 25C, respectively.
After being added for each frame by 5E and 25F, it is sent to the scene change detector 28 and used for scene change determination. Macro block activity ACTmb = a ACTcur + b ACTpred + c ACTmv (21)

The values of the constants a, b, and c are changed for each picture, each macroblock prediction mode (intra, unidirectional prediction, or bidirectional prediction). For example, in the case of intra, prediction is not performed as in the case of an I picture.
= C = 0, and it is considered that the generated code amount is larger than that of the block to be predicted. Therefore, the value of a is increased.
As described above, by performing the activity detection in accordance with the prediction mode or the like, it is possible to estimate the screen complexity more in accordance with the encoding characteristics than in the second embodiment.

The ACTcu added from the image characteristic detector 25 to the scene change detector 28 on a frame-by-frame basis.
The values of r, ACTpred, and ACTmv are sent. In the case of an I-picture, only ACTcur is sent, so that the ratio of the activity of the current picture to be encoded to the activity of the I-picture just encoded is calculated for ACTcur as in the second embodiment. When the ratio of this activity exceeds a certain range, for example, (ratio) <Amin or (ratio)> Amax where (ratio) = (activity of coded image) / (activity of previous image) 0 <Amin If <1, Amax> 1, it is determined that a scene change has occurred between these two images.

On the other hand, in the case of P and B pictures, ACTcur,
Since three values of ACTpred and ACTmv are sent from the image characteristic detector 25, for each of the three values, the activity of the current image to be encoded and the activity of the image of the same picture type encoded immediately before are calculated. Calculate the ratio. And ・ (Ratio of ACTcur) <A1min or (Ratio of ACTcur)
> A1max ・ (Ratio of ACTpred) <A2min or (Ratio of ACTpred)
> A2max • (ACTmv ratio) <A3min or (ACTmv ratio)>
A3max where (ratio) = (activity of coded image) / (activity of previous image) 0 <A1min, A2min, A3min <1, A1max, A2max, A3max,
When any one of> 1 is satisfied, it is determined that a scene change has occurred between these two images.

If it is determined that the possibility of erroneous determination is high with one determination method, the condition may be limited so that it is determined that a scene change has occurred only when all two or all three conditions are satisfied. Good.

The position information of the scene change image is ACTcu
Information indicating which of r, ACTpred, and ACTmv has been determined as a scene change is also sent to the screen complexity calculator 24 at the same time. In addition, the screen complexity calculator 24 determines, based on which activity determination the scene change is determined,
Modify the functions of f1 to f4 in equations (5) to (8).

As described above, in the third embodiment, not only the activity ACTcur of the original image but also the activities ACTpred, A of the error image and the motion vector in the motion compensation prediction.
By using CTmv for scene change determination, the accuracy of scene change determination can be improved. In the third embodiment, the method is applied to the one-pass variable bit rate encoding of the second embodiment. However, the same method is applied to the fixed bit rate encoding of the first embodiment. May be applied.

In the above embodiment, three types of picture coding structures, i.e., I picture, P picture, and B picture as shown in FIG.
There may be only two types such as a picture, an I picture and a B picture. Also, all pictures may be I pictures for which motion compensation prediction is not performed. However, this I
The third embodiment in the case of only pictures is exactly the same as the second embodiment because the input to the image characteristic detecting unit 25 is only the divided moving image.

Further, in the above-described embodiment, scene change determination is separately performed for each of the I, P, and B picture types to determine a scene change image. However, in the case of M> 2, as shown in FIG. 5, when a scene change occurs between P or I pictures sandwiching a B picture, prediction of all of a plurality of B pictures therebetween is less than normal. It may be difficult to hit.

Therefore, as shown in FIG. 5, the P or I picture immediately after the B picture in the order of the input image,
When a P or I picture immediately following a picture and a B picture sandwiched between the immediately preceding P or I pictures are determined to be scene change images, all B pictures in between may be determined to be scene change images.

When all the consecutive B pictures are determined to be scene change images, the B picture “just before” in the first scene change image is regarded as the image immediately before the scene change, and the subsequent B pictures are determined. Also commonly used for pictures. The image following the scene change image is the B picture next to the last scene change image, and the image immediately before the scene change to be compared uses the B picture immediately before the first scene change image.

However, in this case, even when the equations (15) and (16) are used in estimating the screen complexity of the next B picture of the scene change image, the estimation is not performed until the estimation of the screen complexity of the B picture is completed. B just before the first scene change image
Since the activity of the picture and the actually measured screen complexity need to be stored in the screen complexity memory 24M, it is necessary to appropriately switch the image stored in the activity in the screen complexity memory 24M depending on the presence or absence of a scene change and the configuration of the scene change image. is there.

In the second and third embodiments, 1 GOP
The average screen complexity required when calculating the required allocated code amount Rc of each frame was obtained for each coded picture type, but this was not distinguished by the picture type, and the screen complexity of each frame within a certain period was added. Later, the value divided by the number of frames during the period is calculated as the average screen complexity X-ave,
And the estimated screen complexity of the current image Xk (k = i or p or
From b), the required allocated code amount Rc of one GOP may be calculated by the following equation (22).

Required code amount Rc = Rave · (Xk / X-ave) (where k = i or p or b) (22)

[0086]

As described above, according to the present invention, when a moving image is encoded by fixed bit rate control or variable bit rate control of the one-pass method, the amount of generated code of an image in a fixed section in which encoding has been completed. , Average quantization scale, coded image characteristics of the fixed interval and the current image to be coded
(Activity) is detected, a value obtained by performing a predetermined operation on the product of the generated code amount and the average quantization scale is obtained as the actually measured screen complexity, and then, for the immediately preceding image of the same picture type, The presence / absence of a scene change is determined based on the activity ratio of the current image, and the screen complexity of the image to be encoded is multiplied by a predetermined function having the activity ratio as a factor for the scene change image. For the next image of the scene change image, the ratio of the activity of the current image to the image immediately before the scene change image is estimated by multiplying the actually measured screen complexity in the scene change image. By performing code amount allocation at the time of a scene change using the estimated screen complexity, the scene Immediately after the change point of the image such Enji, image deterioration is not generated, without also not occur waste code amount Conversely, it is possible to perform the code amount allocation without the corresponding excess and deficiency.

[Brief description of the drawings]

FIG. 1 is a first embodiment of a moving picture coding apparatus and method according to the present invention.
FIG. 3 is a diagram showing a block configuration of the example of FIG.

FIG. 2 shows a second example of the moving picture coding apparatus and method according to the present invention.
FIG. 3 is a diagram showing a block configuration of the example of FIG.

FIG. 3 is a third embodiment of the moving picture coding apparatus and method according to the present invention;
FIG. 3 is a diagram showing a block configuration of the example of FIG.

FIG. 4 is a diagram showing a block configuration of an embodiment of an image characteristic detector and a scene change detector according to a third embodiment of the present invention.

FIG. 5 is a diagram illustrating an embodiment of a scene change determination for a B picture according to the present invention.

FIG. 6 is a diagram illustrating an example of a coded picture structure.

FIG. 7 is a diagram illustrating a block configuration of a configuration example of a general moving image encoding device.

FIG. 8 is a diagram illustrating a block configuration of a configuration example of a conventional moving picture encoding device.

[Explanation of symbols]

Reference Signs List 11 subtracter 12 DCT unit (orthogonal transformer) 13 quantizer 14 code amount controller 15 variable length encoder 16 buffer 17 inverse quantizer 18 IDCT unit 19 motion compensation predictor 20 adder 21 frame memory 22 average quantum Scaled scale detector 23 Generated code amount detector 24 Screen complexity calculator 24M Screen complexity memory 25 Image characteristic detector 25A ACTcur detector 25B ACTpred detector 25C ACTmv detector 25D, 25E, 25F Accumulator 25G Picture activity calculator 28 Scene Change Detector 28A SCcur Detector 28B SCpred Detector 28C SCmv Detector 28D Scene Change Detector 29 Average Screen Complexity Calculator ACTcur Original Image Activity ACTi, ACTp, ACTb Current Image Activity ACTi-p, ACTp-p , ACTb-p Activates the previously coded image of the same picture type. Activity ACTmb Macroblock activity ACTmv Motion vector characteristics ACTp-pold, ACTb-pold Activity of image subject to measured screen complexity ACTpred Error image activity R Code amount Rave Average allocated code amount Rc Average allocated code amount Xi, Xp, Xb Screen complexity of current image Xi-ave, Xp-ave, Xb-ave Average screen complexity Xi-p, Xp-p, Xb-p Actual measurement screen complexity of each frame Xp-pold, Xb-pold Actual measurement screen Complexity

Claims (12)

[Claims] 1. A moving picture coding apparatus for coding an input moving picture by including a motion compensation predicting means, an orthogonal transform means, a quantizing means, and a variable length coding means, comprising the steps of: Means for detecting a generated code amount; means for detecting an average quantization scale of each image of the input moving image; and at least the input among the input moving image and the motion-compensated predicted image generated by the motion-compensated predicting means. Means for detecting a coded image characteristic of a moving image, means for detecting scene change image information from the coded image characteristic detected by the means for detecting the coded image characteristic, and means for detecting the generated code amount From the detected generated code amount and the average quantization scale detected by the means for detecting the average quantization scale, the actually measured screen complexity of the past image is calculated. From the calculated actual screen complexity of the past image, the coded image characteristic detected by the coded image characteristic detecting means, and the scene change image information detected by the scene change image information detecting means, Means for calculating the estimated screen complexity of the image, and means for calculating the estimated screen complexity of the past image calculated by the means for calculating the actually measured screen complexity, or the estimated screen complexity of the current image. The calculated estimated screen complexity of the current image, the generated code amount detected by the means for detecting the generated code amount, the average quantization scale detected by the means for detecting the average quantization scale, the coding From the encoded image characteristics detected by the image characteristic detecting means, the allocated code amount of the next image to be encoded is determined, and the allocated code amount is determined. Video encoding apparatus characterized by comprising a means for determining the quantization scale of the image to be the next coded from the amount. 2. The moving picture coding apparatus according to claim 1, wherein the means for calculating the estimated screen complexity of the current image includes: a coded image characteristic of the detected current image; Multiplying the measured screen complexity in the immediately preceding picture by a predetermined function having a factor of a ratio with a coded image characteristic detected in the picture immediately before the picture type (I picture, P picture, B picture). A moving image encoding apparatus for calculating the estimated screen complexity by using the method. 3. The video encoding apparatus according to claim 1, wherein the current image is determined to be a scene change image or an image next to the scene change image based on the detected scene change image information. Changes a predetermined function whose factor is a ratio with the coded image characteristic. If the current image is determined to be the next image of the scene change image, the same picture type (I A moving picture encoding apparatus for changing a picture immediately before a picture, a P picture, and a B picture into a picture preceding a scene change picture of the same picture type as the current picture. 4. The moving picture coding apparatus according to claim 1, wherein the screen complexity used in said means for determining the assigned code amount of the picture to be coded next is at least said current picture being said detected. A moving picture encoding apparatus characterized in that, when it is determined based on the obtained scene change picture information that the picture is a scene change picture or an image next to the scene change picture, the estimated screen complexity of the current picture is used. 5. The moving picture coding apparatus according to claim 1, wherein an image to be coded next based on said generated code amount and said measured screen complexity or said estimated screen complexity. Means for determining the allocated code amount, the ratio of the estimated screen complexity to the average value of the measured screen complexity in a certain period, the average allocated code amount is multiplied to determine the allocated code amount. Video encoding device. 6. The moving picture coding apparatus according to claim 1, wherein said means for detecting a coded picture characteristic of the input moving picture or the motion-compensated predicted picture comprises: A picture type (I picture, P picture, B picture) comprising: means for detecting an image property of an error image of a motion compensated prediction image; and means for detecting a motion vector property in motion compensation prediction. )
Constants for different and prediction modes, the image characteristics of the detected input moving image, the image characteristics of the error image of the motion compensated prediction image,
By multiplying and adding each of the characteristic values of the motion vector characteristic in the motion compensation prediction, the encoded image characteristic is determined, and the means for detecting the scene change image information includes the current image and the current image. Between the immediately preceding pictures of the same picture type (I picture, P picture, B picture), the picture characteristics of the input moving picture detected by the means for detecting the coded picture properties, the error of the motion compensated prediction picture Image characteristics of the image, and calculate the ratio of the respective characteristic values of the motion vector characteristics in the motion compensation prediction, when the ratio of the characteristic values exceeds a certain range determined for each characteristic value, the current image Is determined to be a scene change image, and the next image of the same picture type (I picture, P picture, B picture) as the scene change image is next to the scene change image. A moving picture coding apparatus characterized in that it is defined as an image. 7. A moving picture coding method for coding an input moving picture by including a motion compensation prediction step, an orthogonal transformation step, a quantization step, and a variable length coding step, comprising the steps of: Detecting a generated code amount; detecting an average quantization scale of each image of the input moving image; at least the input of the motion compensated prediction image generated by the input moving image and the motion compensation prediction step Detecting an encoded image characteristic of the moving image; detecting scene change image information from the encoded image characteristic detected by the detecting the encoded image characteristic; and detecting the generated code amount. The detected generated code amount and the average quantization scale detected by the step of detecting the average quantization scale. Calculating the actually measured screen complexity of the past image from the previous image, calculating the actually measured screen complexity of the past image, the coded image characteristic detected by the step of detecting the coded image characteristic, and the scene change. Calculating the estimated screen complexity of the current image from the scene change image information detected by the image information detecting step; and the measured screen complexity of the past image calculated by calculating the measured screen complexity Or the estimated screen complexity of the current image calculated by the step of calculating the estimated screen complexity of the current image, the generated code amount detected by the step of detecting the generated code amount, the average quantization scale Detecting the average quantization scale detected by the step of detecting the encoded image characteristic, From out encoded image characteristics,
Determining an assigned code amount of an image to be encoded next, and determining a quantization scale of the image to be encoded next from the assigned code amount. 8. The moving picture coding method according to claim 7, wherein the step of calculating the estimated screen complexity of the current image includes the steps of: determining a coded image characteristic of the detected current image; Picture type (I picture, P picture, B picture) multiplying the measured screen complexity in the immediately preceding image by a predetermined function having a factor of a ratio with an encoded image characteristic detected in the immediately preceding image. A video encoding method for calculating the estimated screen complexity. 9. The moving picture encoding method according to claim 7, wherein the current image is determined to be a scene change image or an image next to the scene change image based on the detected scene change image information. Changes a predetermined function having a ratio to the encoded image characteristic as a factor, and when the current image is determined to be the next image of the scene change image, the same picture type as the current image
A moving picture coding method, wherein an image immediately before (I picture, P picture, B picture) is changed to a picture before a scene change picture of the same picture type as the current picture. 10. The moving picture coding method according to claim 7, wherein the screen complexity used in the step of determining the assigned code amount of the picture to be coded next is at least that the current picture is the detected picture. A moving image encoding method, characterized in that when it is determined based on the obtained scene change image information that the image is a scene change image or an image next to the scene change image, the estimated screen complexity of the current image is used. 11. The moving picture coding method according to claim 7, wherein an image to be coded next based on the generated code amount and the measured screen complexity or the estimated screen complexity. Determining the allocated code amount of, the ratio of the estimated screen complexity to the average value of the measured screen complexity for a certain period, the average allocated code amount is determined by multiplying the average allocated code amount is determined. Moving image encoding method. 12. The moving picture coding method according to claim 7, wherein the step of detecting a coded picture characteristic of the input moving picture or the motion-compensated predicted picture comprises the step of: Detecting an image characteristic of an error image of the motion-compensated prediction image, and detecting a motion vector characteristic in the motion-compensated prediction. The picture type (I picture, P picture, B picture) )
Constants for different and prediction modes, the image characteristics of the detected input moving image, the image characteristics of the error image of the motion compensated prediction image,
By multiplying and adding each of the characteristic values of the motion vector characteristic in the motion compensation prediction, the encoded image characteristic is determined, and the step of detecting the scene change image information includes: The image characteristics of the input moving image detected by the step of detecting the coded image characteristics between the immediately preceding images of the same picture type (I picture, P picture, B picture), errors in the motion compensated prediction image The image characteristics of the image, and calculate the ratio of the respective characteristic values of the motion vector characteristics in the motion compensation prediction, if the ratio of the characteristic values exceeds a certain range defined for each characteristic value, the current image Is determined as a scene change image, and the next image of the same picture type (I picture, P picture, B picture) as the scene change image is determined as the scene change image. A moving picture coding method characterized in that it is determined as the next picture of an image.