JPH06105299A - Dynamic image compressor - Google Patents

Abstract

PURPOSE:To suppress the dispersion of moving vectors obtained by a movement prediction, to improve an entire encoding efficiency, and to improve a picture quality by selecting the moving vector whose efficiency is the most effective at the time of actuallyencoding among the plural detected candidate moving vectors. CONSTITUTION:In a dynamic image compressor accompanied with a time prediction, generally the difference of a picture at a certain point of time is data to be compressed by the prediction from the picture in the past or future, and one picture 10 is divided into several partial pictures 11, and a prediction processing including a movement compensation is operated to each partial picture 11. In this case, the dynamic image compressor detects the several candidate moving vectors of the partial pictures 11 with a predicted error, calculates an error coefficient reflecting a code length from the predicted error, calculates a vector encoding coefficient from the encoded moving vector, and decides the moving vector to be adopted based on each coefficient.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture compression apparatus used for moving picture compression processing and the like, and more particularly to a moving picture compression apparatus with prediction in the time axis direction.

[0002]

2. Description of the Related Art As an international standard for image compression, JPEG (Jo
int Photograghic Expert Group) and MPEG (Moving
There is a Picture Expert Group).

MPEG is MPEGI, MPEGII, M
A three-level draft of PEGIII is under consideration. MP
EGI aims to compress a moving image that can be transmitted through a communication line of 1.5 Mbps, and is considered to be used mainly in videophones and videoconferences. MPEG I
Then, the current NTSC format video image is 320 x 24
It is treated as a resolution of 0 pixel, and data of only one field is used out of two fields constituting one frame. In MPEGII, compression is a goal that can be transmitted through a 10 Mbps communication line, and moving image transmission by ISDN and digital video are targets. And
MPEGIII is targeted for next-generation televisions such as HDTV.

The feature of MPEG is that DCT (Discrete Cos
ine Transform: Discrete Cosine Transform) is used to perform inter-frame prediction processing for compression in the time axis direction in addition to still image compression. Random access of frames and fast-forwarding are prerequisites for moving image compression. It has been mentioned that playback and rewind playback (reverse direction) can be performed. Therefore, inter-frame prediction in MPEG employs both forward and backward directions. MPEG
However, basically, MC (motion compensation) + DCT is used. The block size for motion compensation is 16 × 16.
(However, there is also an 8 × 8 mode), DCT is performed on 8 × 8 blocks. Also, this motion compensation is performed with 1/2 pixel precision. Motion compensation with 1/2 pixel accuracy is performed not only by checking the position shifted in pixel units on the reference frame used for prediction, but also by generating the position between pixels by interpolation and performing matching. In a moving image compression apparatus that involves prediction in the time direction, motion-based prediction is performed in order to reduce deterioration of prediction efficiency due to camera PAN and movement of a subject. In this motion compensation, the motion vector of the target area is detected between the target frame and the reference frame, and the position shifted by the motion vector in the reference frame is used as the reference pixel, and this is used as the prediction value, and the difference from the target pixel (prediction error) Is a method of transmitting. For example, in motion compensation prediction, as shown in FIG. 9, the motion of the moving object is predicted based on the motion vector of the prediction source image, and the motion is compensated for in the original image. 16x1 motion compensation
This is a method of searching for a portion having the smallest difference in the vicinity of the position of the block of the previous image in units of 6 pixels, and taking the difference with it to further reduce the data to be transmitted. As a detecting means, as shown in FIG. 10, generally, a method of searching within a certain range from the original position of a partial image to be motion-compensated and selecting a position with the smallest error is adopted.

In addition, an ordinary moving picture compression apparatus (CCITT H.261 or MPEG.Vid with prediction in the time direction) is used.
eo, etc.), when the generated motion vector is encoded, the difference from the motion vector of the partial image in the vicinity thereof (usually the partial image processed immediately before) is calculated, and only the difference is encoded. (Described later in FIG. 2).

[0006]

However, in such a conventional moving picture compression apparatus, the difference between the motion vectors of the partial image processed immediately before is calculated and moved to the place where the difference is the smallest. Since the configuration is such that the PAN or the like is encoded, the difference is concentrated near 0 when the motion of the PAN or the like is directed in a certain direction. When the motion vector varies depending on the partial image, there is a drawback that it is further dispersed by taking the difference and consequently the coding efficiency is reduced. For example, if noise is present on the moved screen, the noise will make the motion vector detection inaccurate and reduce the coding efficiency.

Therefore, an object of the present invention is to provide a moving picture compression apparatus capable of suppressing the variation of the motion vector obtained by the motion estimation and improving the coding efficiency.

[0008]

The invention according to claim 1 is
In order to achieve the above object, in a moving picture compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in the time axis direction, an input screen is divided into predetermined blocks and motion vectors are calculated. A moving image compression apparatus comprising a motion compensating unit prepared for each block and performing inter-frame prediction by shifting by a motion vector from a previous reproduced image, wherein a plurality of motion vector candidates are detected and Of the motion vectors, the most efficient motion vector is selected when actually encoding.

According to a second aspect of the present invention, in a moving image compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, an input screen is divided into predetermined blocks. A motion vector compression apparatus is provided with a motion compensation unit that prepares a motion vector for each block and shifts the motion vector from the previous reproduced image by the motion vector to perform inter-frame prediction. Vector detection means, prediction error detection means for detecting the prediction error of a plurality of candidate motion vectors detected by the motion vector detection means, and an error coefficient for the minimum prediction error for the prediction error detected by the prediction error detection means. Error coefficient calculation means for calculating and actual encoding based on the error coefficient calculated by the error coefficient calculation means And a motion vector determining means for determining the most efficient motion vector can.

According to a third aspect of the invention, in a moving image compression apparatus with time prediction in which image data is converted to a frequency axis by an orthogonal transformation means and compressed in a time axis direction, an input screen is divided into predetermined blocks. A motion vector compression apparatus is provided with a motion compensation unit that prepares a motion vector for each block and shifts the motion vector from the previous reproduced image by the motion vector to perform inter-frame prediction. Vector detection means, prediction error detection means for detecting the prediction error of the plurality of candidate motion vectors detected by the motion vector detection means, and actually for the plurality of candidate motion vectors detected by the motion vector detection means Coded motion vector calculation means for calculating a motion vector to be coded, and calculation by the coded motion vector calculation means Coding coefficient calculation means for calculating a code length when coded based on the coded motion vector as a code length vector coding coefficient, and actually based on the coding coefficient calculated by the coding coefficient calculation means. And a motion vector determining means for determining the most efficient motion vector when encoding.

According to a fourth aspect of the invention, in a moving image compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, an input screen is divided into predetermined blocks. A motion vector compression apparatus is provided with a motion compensation unit that prepares a motion vector for each block and shifts the motion vector from the previous reproduced image by the motion vector to perform inter-frame prediction. Vector detection means, prediction error detection means for detecting the prediction error of the motion vectors of the plurality of candidates detected by the motion vector detection means, and encoding based on the motion vectors of the plurality of candidates detected by the motion vector detection means By the coding coefficient calculation means for calculating the code length at this time as the code length vector coding coefficient, and the coding coefficient calculation means And a motion vector determining means for determining the most efficient motion vector when actually encoding based on the issued coding coefficients.

According to a fifth aspect of the present invention, in a moving image compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, an input screen is divided into predetermined blocks. A motion vector compression apparatus is provided with a motion compensation unit that prepares a motion vector for each block and shifts the motion vector from the previous reproduced image by the motion vector to perform inter-frame prediction. Vector detection means, prediction error detection means for detecting the prediction error of a plurality of candidate motion vectors detected by the motion vector detection means, and an error coefficient for the minimum prediction error for the prediction error detected by the prediction error detection means. An error coefficient calculating means for calculating and a plurality of candidate motion vectors detected by the motion vector detecting means A coded motion vector calculation means for calculating a motion vector to be coded at the time, and a code length when coded based on the coded motion vector calculated by the coded motion vector calculation means as a vector coding coefficient. A coding coefficient calculating means for calculating, and the most efficient motion vector when actually coding based on the error coefficient calculated by the error coefficient calculating means and the coding coefficient calculated by the coding coefficient calculating means. And a motion vector determining means for determining.

The orthogonal transform means may be a discrete cosine transform (DCT) for performing a discrete cosine transform (DCT), for example, as described in claim 6, and the temporal prediction is described in claim 7, for example. As described above, inter-frame prediction processing for compression in the time axis direction may be performed.

[0014]

According to the invention described in claims 1, 2, 3, 4, 5, 6 and 7, first, the motion vector detecting means detects a plurality of motion vectors for each cut-out image block for motion compensation. Then, the prediction error detection means detects the prediction errors of the plurality of candidate motion vectors, and the error coefficient calculation means calculates an error coefficient for the minimum prediction error for the detected prediction errors.

Further, when the motion vector actually calculated by the coding motion vector calculating means is calculated with respect to the plurality of candidate motion vectors, and the coding coefficient calculating means codes based on the coded motion vector. The code length of is calculated as the vector coding coefficient.

Then, based on the error coefficient and the coding coefficient, the motion vector determining means determines the most efficient motion vector when actually coding.

Therefore, the variation of the motion vector obtained by the motion estimation can be suppressed, and the coding efficiency can be improved.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings.

Description of Principle First, the basic concept of the present invention will be described.

Generally, in a moving image compression apparatus involving temporal prediction, an image at a certain time point has its difference as data to be compressed by prediction from an image in the past or in the future. As shown in FIG. 1, one image 10 is divided into several partial images 11, and a prediction process including motion compensation is separately performed for each partial image 11. The motion vector thus obtained is coded by taking the difference from the motion vector of the previous partial image. For example, when one piece of image data is encoded in the order from the upper left partial image to the right as shown in FIG. 2, the motion vector of each partial image is taken as a difference from the motion vector on the left side thereof. It is encoded. Here, when the motion vector of each partial image is A, B, C, D, ..., The encoded values are A, BA, CB, DC ,.
Conventionally, since the difference between the motion vectors of the partial image processed one before in this way is taken and encoded, noise is reduced.
In some cases, the motion vector detection may be inaccurate when an elephant rides on.

In view of this, the present invention takes out some candidates when generating a motion vector, makes a trade-off between the prediction error and the variation of the motion vector, and determines the motion vector based on the result. That is, even if the prediction error is the smallest, it may be affected by noise, etc.Therefore, some candidates where the prediction error is small are listed, and the motion vector indicating how much motion has occurred is also given. At the same time, it is detected as data, and the optimum motion vector is determined based on the prediction error, the motion vector, and the coding coefficient indicating the code length when coded. As described above, according to the present invention, not only the prediction error is targeted in the motion vector detection at the time of motion compensation, but also the code amount of the motion vector is targeted, so that the overall coding efficiency is improved and the image quality is improved. To be able to Embodiment FIG. 3 to FIG. 8 are views showing an embodiment of a moving picture compression apparatus according to the present invention.

First, the structure will be described. FIG. 3 is a block diagram of a moving image compression apparatus. In this figure, an encoder of the moving image compression apparatus outputs an image mode, a prediction mode, a motion vector and various control signals to control the entire system. The controller 30, an image memory 31 for storing image data to be data-compressed, a subtracter 32 for subtracting the prediction result of the motion-compensated inter-frame prediction process from the image data read from the image memory 31, and a subtractor 32 for subtraction. DC which performs DCT operation on the image data while outputting the image data to the controller 30.
The T calculation unit 33, the quantization unit 34 that quantizes the output data of the DCT calculation within a certain error range according to the quantization width determined by the controller 30, and the image data quantized by the quantization unit 34. On the other hand, in addition to image data, various block attribute signals are subjected to variable length coding, and then multiplexed into a code string of a predetermined data structure, which is a VLC (Variable Length Co).
de: variable length coding) 35, a buffer 36 that smoothes fluctuating information generation to a constant rate, and a motion-compensated inter-frame prediction unit 37 that performs motion-compensated prediction based on periodic intra-frame coded frames. ,.

The motion-compensated inter-frame prediction unit 37 dequantizes the image data quantized by the quantization unit 34 and the dequantization unit 38, and the dequantization unit 38 restores the unquantized image data. Inverse DCT (IDC
T) According to the image mode and prediction mode from the controller 30, the IDCT calculation unit 39 that performs the calculation, the adder 40 that adds motion compensation to the data returned to the image data before the DCT processing by the IDCT calculation unit 39. It is composed of switches 41, 42 and 43 for switching signal paths, and predictors 44 and 45 for performing motion compensation prediction based on the motion vector calculated by the controller 30 (see FIG. 4).

Next, the operation of this embodiment will be described.

FIG. 4 is a flowchart showing a motion vector determining process of the moving picture compression apparatus.

First, in step S1, a number of motion vector candidates of a partial image are detected together with a prediction error, and in step S2 a prediction error reflecting the coding length is calculated from the image prediction error. That is, here, in the case of predicting a partial image, in addition to the motion vector with the minimum prediction error, the prediction error falls within a certain range (for example, within 5% of the predicted minimum prediction error value of the partial image). Motion vectors are listed as candidates. At that time, a value (herein referred to as an error coefficient) indicating how much the error is with respect to the minimum prediction error value is calculated according to the prediction error, as shown in FIG. Here, the prediction error of the partial image may be, for example, the sum of squares of the prediction errors (differences) of all the pixels in the partial image. Further, the error coefficient can be easily calculated from the minimum prediction error as shown in (Equation 1) as the simplest method, but in reality, the error coefficient is encoded according to the method of encoding the prediction error. It is better to use a calculation method that better reflects the code length of time.

(Prediction error ÷ Minimum prediction error−1) × 70 + 1 = Error coefficient (Equation 1) More specifically, a candidate to be actually used is selected from prediction errors of a plurality of candidates. In the example, a coefficient called an error coefficient is introduced as the method. This error coefficient is a coefficient that relatively indicates how much error is present from the minimum prediction error with reference to the smallest prediction error among the candidates, and is calculated by, for example, (Equation 1).

FIG. 5 shows a partial image (n) generated at the time of prediction.
Is a diagram showing motion vector candidates of, for example, the motion vector is X-axis “−10”, Y-axis “+7” and the prediction error (total difference between the original image and the image to be encoded)
Is "1800", the motion vector is X-axis "-7", the Y-axis is "+5" and the prediction error is "1827", and the motion vector is X-axis "3". , Y axis is “−24” and the prediction error is “1”.
In addition to the motion vector having the smallest prediction error, such as the motion vector candidate of “833”, an error coefficient indicating the degree of error with respect to the minimum prediction error value is obtained according to (Equation 1). .

For example, the motion vector X axis "-10", Y
The error coefficient of the axis "+7" is (1800/18
00-1) × 70-1 = 1. Note that “70” in (Equation 1) is a gain adjustment gain.

When the motion vectors of a plurality of candidates and their error coefficients are obtained in this way, the process moves to the process of calculating the motion vector actually encoded.

Returning to the flow of FIG. 4, in step S3, a value to be actually encoded is calculated from the motion vectors of the number candidates, and in step S4, a vector coding coefficient (how much to reflect the motion vector coding length) is calculated. The coefficient representing whether or not the code length is calculated. That is, since the motion vector is coded by taking the difference from the adjacent block, the coded motion vector is more important, and the coded motion vector is first calculated.

FIG. 6 is a diagram showing the motion vector to be coded and the vector coding coefficient calculated for the motion vector candidate of the partial image (n). The temporal distance between the prediction source image and the current image is shown in FIG. This is an example when it is 2 (Frames).
For example, in this figure, when the motion vector of the immediately preceding left adjacent block is (x, y) = (− 7, +2), the motion vector encoded is “−3” on the X axis and “3” on the Y axis. +5 "
Becomes The vector coding coefficient is a coefficient indicating how much code length (the number of bits) is obtained when the motion vector to be actually coded is coded, and the coding length is the prediction source image and the current image. Since it often changes depending on the temporal distance between the two, the calculation is performed as shown in FIG. 7, for example. In this case, since it depends on how the moving picture compression system encodes, it is preferable to use a calculation method suitable for the system.

Based on the error coefficient and the vector coding coefficient thus obtained, the most efficient one is selected at the time of actual coding in step S5 of FIG. For example, as the simplest method, the one with the smallest sum of the coefficients may be selected. FIG. 8 shows an example in which the second candidate having the smallest sum of the coefficients is selected. That is, the first candidate can be selected only by focusing on the error coefficient, but the optimum candidate is selected by selecting the one with the smallest sum of the coefficients including the vector coding coefficient. Become. Here, the second candidate is selected as a candidate to be actually used.

If it is desired to obtain more complete precision, there is also a method of actually encoding each coefficient to obtain the code length.

As described above, the moving picture compression apparatus according to the present embodiment detects a number of motion vector candidates of a partial image together with a prediction error, calculates an error coefficient reflecting the code length from the prediction error, and calculates the code. Since the vector coding coefficient is calculated from each motion vector and the motion vector to be adopted is determined based on each coefficient, it is possible to suppress the variation of the motion vector obtained by the motion prediction, and to encode the entire coding. The efficiency can be improved and the image quality can be improved. That is, the present moving image compression apparatus detects a plurality of motion vectors, and selects an optimum motion vector from the candidates to increase coding efficiency. Motion vector detection during motion compensation In (1), not only the prediction error but also the code amount of the motion vector is targeted, so that the overall coding efficiency can be improved and the image quality can be improved. Further, in the present embodiment, the error coefficient reflecting the coding length, the coding motion vector actually to be coded, the vector coding coefficient are calculated, and the motion vector to be adopted is determined based on each coefficient. However, any configuration may be used as long as it detects a plurality of motion vectors and selects the most efficient motion vector from the plurality of detected motion vectors when actually encoding. . For example, a mode may be adopted in which the motion vector is selected from the candidates without calculating the error coefficient, the coded motion vector, or the vector coded coefficient. However, if the configuration as in this embodiment is adopted, it becomes possible to adopt the optimum motion vector.

In this embodiment, the moving picture compression device is set to MP.
This is an example of application to a moving image compression apparatus based on the EG algorithm, but of course the present invention is not limited to this, and can be applied to any apparatus as long as it determines the edge of a moving body and raises the importance of encoding. Needless to say.

Further, in this embodiment, the conversion coding method is set to D.
Although CT is applied, the present invention is not limited to this DCT method, and is applied to, for example, a moving image compression apparatus using Hadamard transform, Harr transform, slant transform (slant transform), symmetric sine transform, or the like. be able to.

Further, it is needless to say that the number and types of circuits and members constituting the above moving picture compression apparatus are not limited to those in the above-mentioned embodiment, but may be realized by software (for example, C language). .

[0039]

According to the first, second, third, fourth, fifth, sixth and seventh aspects of the present invention, the input screen is divided into predetermined blocks and the motion vector is prepared for each block. A moving image compression apparatus including motion compensation means for performing inter-frame prediction by shifting by a motion vector from
A plurality of candidates for the motion vector are detected, and the most efficient motion vector when actually encoding is selected from among the plurality of detected motion vectors. It is possible to suppress, the overall coding efficiency can be increased, and the image quality can be improved.

[Brief description of drawings]

FIG. 1 is a diagram showing a partial image of a moving image compression apparatus.

FIG. 2 is a diagram showing encoded motion vector values of a moving picture compression apparatus.

FIG. 3 is a diagram showing a block configuration of a moving image compression apparatus.

FIG. 4 is a flowchart showing an operation of the moving image compression apparatus.

FIG. 5 is a diagram showing motion vector candidates generated at the time of prediction by the moving image compression apparatus.

[Fig. 6] Fig. 6 is a diagram illustrating a motion vector and a vector coding coefficient that are coded by the moving image compression apparatus.

FIG. 7 is an example of calculation of vector coding coefficients of a moving image compression apparatus.

FIG. 8 is an example of selection of candidates to be actually used by the moving image compression apparatus.

FIG. 9 is a diagram showing motion-compensated prediction of a moving image compression apparatus.

FIG. 10 is a diagram showing a motion vector search range of the moving image compression apparatus.

[Explanation of symbols]

 30 controller 31 image memory 32 subtractor 33 DCT operation unit 34 quantization unit 35 VLC 36 buffer 37 motion compensation inter-frame prediction unit 38 dequantization unit 39 IDCT operation unit 40 adder 41, 42, 43 switch 44, 45 predictor

Claims (7)

[Claims] 1. A moving picture compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, in which an input screen is divided into predetermined blocks and motion vectors are divided into blocks. A moving image compression apparatus, which is provided for each of the plurality of motion vector detection means, detects a plurality of motion vector candidates and detects a plurality of detected motion A moving image compression apparatus characterized in that, of the vectors, the most efficient motion vector is selected when actually encoding. 2. In a moving picture compression apparatus with time prediction for converting image data into a frequency axis by an orthogonal transformation means and compressing it in the time axis direction, an input screen is divided into predetermined blocks and motion vectors are divided into blocks. A moving image compression apparatus that is provided for each of the moving images, and that includes a motion compensation unit that performs inter-frame prediction by shifting by a motion vector from a previously reproduced image, the motion vector detecting unit that detects a plurality of motion vector candidates, Prediction error detection means for detecting prediction errors of the plurality of candidate motion vectors detected by the motion vector detection means, and error coefficient calculation means for calculating an error coefficient for the minimum prediction error of the prediction error detected by the prediction error detection means And the most efficient when actually encoding based on the error coefficient calculated by the error coefficient calculating means. Video compression apparatus characterized by comprising a motion vector determining means, for determining a can vector. 3. A moving picture compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, wherein an input screen is divided into predetermined blocks and motion vectors are divided into blocks. A moving image compression apparatus that is provided for each of the moving images, and that includes a motion compensation unit that performs inter-frame prediction by shifting by a motion vector from a previously reproduced image, the motion vector detecting unit that detects a plurality of motion vector candidates, Prediction error detection means for detecting the prediction error of the motion vectors of the plurality of candidates detected by the motion vector detection means, and a motion vector actually encoded with respect to the motion vectors of the plurality of candidates detected by the motion vector detection means And a coded motion vector calculated by the coded motion vector calculation means. Coding coefficient calculation means for calculating the code length when coded based on Toll as a code length vector coding coefficient, and when actually coding based on the coding coefficient calculated by the coding coefficient calculation means A moving image compression apparatus comprising: a motion vector determining unit that determines the most efficient motion vector. 4. A moving picture compression apparatus with time prediction in which image data is converted into a frequency axis by an orthogonal transformation means and compressed in a time axis direction, wherein an input screen is divided into predetermined blocks and motion vectors are divided into blocks. A moving image compression apparatus that is provided for each of the moving images, and that includes a motion compensation unit that performs inter-frame prediction by shifting by a motion vector from a previously reproduced image, the motion vector detecting unit that detects a plurality of motion vector candidates, Prediction error detection means for detecting the prediction error of the motion vectors of the plurality of candidates detected by the motion vector detection means, and the code length when encoded based on the motion vectors of the plurality of candidates detected by the motion vector detection means Coding coefficient calculation means for calculating as a coding length vector coding coefficient, and a coding coefficient calculated by the coding coefficient calculation means Indeed video compression apparatus characterized by comprising a motion vector determining means, for determining the most efficient motion vector when coding based on. 5. A moving picture compression apparatus with time prediction for converting image data to a frequency axis by an orthogonal transformation means and compressing it in the time axis direction by dividing an input screen into predetermined blocks and dividing motion vectors into blocks. A moving image compression apparatus that is provided for each of the moving images, and that includes a motion compensation unit that performs inter-frame prediction by shifting by a motion vector from a previously reproduced image, the motion vector detecting unit that detects a plurality of motion vector candidates, Prediction error detection means for detecting prediction errors of the plurality of candidate motion vectors detected by the motion vector detection means, and error coefficient calculation means for calculating an error coefficient for the minimum prediction error of the prediction error detected by the prediction error detection means And the plurality of candidate motion vectors detected by the motion vector detecting means are actually encoded. Coding vector calculation means for calculating a coding vector, and coding coefficient calculation for calculating a code length when coded based on the coding motion vector calculated by the coding motion vector calculation means as a vector coding coefficient. Means, and a motion vector determination means for determining the most efficient motion vector when actually encoding based on the error coefficient calculated by the error coefficient calculation means and the coding coefficient calculated by the coding coefficient calculation means. A moving image compression apparatus comprising: 6. The orthogonal transform means is a discrete cosine transform for performing a discrete cosine transform (DCT), any one of claim 1, claim 2, claim 3, claim 4, and claim 5. The moving image compression apparatus according to claim 1. 7. The temporal prediction is an interframe prediction process for compression in a time-axis direction, claim 1, claim 2, claim 3, claim 4, or claim. 5. The moving image compression apparatus according to any one of 5 above.