JPH07203434A - Image transmitter - Google Patents

Abstract

PURPOSE:To improve the transmission efficiency by applying efficient coding processing to a picture in an important range designated by the operator while improving the image quality on an emphasized image in the case of coding. CONSTITUTION:When the operator selects the emphasized range designation mode by using an operation section 11, a system control section 12 receiving the mode selection notice gives a command of an emphasized range decided at first to an emphasized range designation section 18, which gives information representing the emphasized range designation to a graphic generating section 17. An image coding section applies coding to picture data within the emphasized range with higher picture quality than the image data of other range.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image transmitting device,
In particular, the present invention relates to an image transmission device that transmits and receives video and audio via a digital communication line.

[0002]

2. Description of the Related Art In the case of a conventional analog telephone line, only voice can be transmitted and data can be transmitted only at a low speed.

However, in recent years, along with the progress of communication technology, semiconductor technology, and optical technology, digital lines have been established, and high-speed and large-capacity data transmission has become possible.

In particular, a characteristic of digital transmission is that the same level of quality is maintained without deterioration of quality due to transmission.
Since it is possible to integrate the media without requiring a transmission path according to the characteristics of the transmission data media, it has become possible to transmit data between the composite media terminals. Therefore, a telephone terminal has emerged which transmits not only the conventional audio-only transmission but also the video transmission.

Under these circumstances, international standardization by CCITT and the like is in progress to enable mutual communication between different compound terminals, and AV (Audio) such as a videophone using a digital line and a delevi conference system is being developed.
As a Visual service, service specifications, protocol specifications, and multimedia multiplexing frame configuration specifications for AV services are CCITT Recommendation (or Draft) H.264. 32
0, H. 242, H.M. It has been announced as 221 etc.

H. In 221, from 64 Kbps to 192
Exchange of frame structure and terminal capability in AV services up to 0 Kbps and FAS (Frame) of communication mode
Alignment Signal), BAS (Bit
Allocation Signal) code allocation is defined.

H. 242 defines a protocol such as capability exchange and communication mode switching between AV terminals using BAS, and H320 defines a system aspect of AV services in general.

In the above recommendation (or draft recommendation),
After setting up end-to-end physical connection and establishing in-channel AFS synchronization, in-channel BA
A method for performing multimedia communication of images, voices, data, etc. between terminals is defined by procedures such as a terminal capability exchange sequence using S and a mode switching sequence by specifying a communication mode.

However, which communication mode is to be used within the range of the capability of each terminal to change its own terminal capability according to the situation or the exchanged capability is outside the stipulated range.

The information transfer rate of each medium in multimedia communication is determined by designating the voice coding method for voice information, and the presence or absence of use of the data information and the transfer rate when it is used. The transfer rate of the image information is obtained by subtracting the transfer rate of the voice information and the transfer rate of the data information from the determined and set information transfer rate of the entire communication path.

[0011]

However, according to the above-mentioned conventional TV telephone apparatus, in the case of the present encoding method, the characteristic of the image in the block is checked in a certain block unit, and the characteristic is determined. The corresponding encoding control method is implemented. Alternatively, encoding control is also performed on the assumption that a person is in the center and the surroundings are the background, assuming a pattern of the conference such as background or person to some extent.

Therefore, in the former case, the judgment is simply made based on the characteristics of the image in the block, so that it becomes a very microscopic view of the image, and it is judged whether or not the image of the block is an important block on the entire screen. It's hard to do. In particular, if the amount of data generated in one frame is limited to an arbitrary value, it is assumed that the importance of the entire screen is the same and the amount of image data is suppressed to a uniform amount. There is a problem in that the quality of an image that is really important cannot be improved due to the processing.

Further, in the latter case, efficient coding is possible when the condition is fixed and the condition is met, but conversely, when the condition is not met, conversely, There is a problem that the encoding control hinders the efficiency improvement and reduces the efficiency more than usual. Further, the above conditions do not always reflect the intention of the operator, so that there is a problem that the requirements of the operator are not satisfied.

Therefore, an object of the present invention is to provide an image transmitting apparatus which solves the above problems.

[0015]

SUMMARY OF THE INVENTION According to the present invention, a display means for displaying a transmission image and information indicating the important position of the image are displayed on the displayed transmission image, and the setting is made to interactively set the important position. Means and encoding means for encoding the image data in the important position of the image to improve the image quality compared to the image data at other positions.

Furthermore, the present invention is preferably characterized by further comprising a designation means for designating at least one important point position of the image.

More preferably, the designation means designates the emphasis position of the image at a plurality of levels.

More preferably, the designating means designates the emphasis position of the image in units of minimum image blocks for coding.

More preferably, the encoding means makes the quantization step width of the image data at the important position of the image smaller than the quantization step width at the positions other than the important position.

More preferably, the encoding means makes the threshold value of the quantized data of the image data at the important position of the image smaller than the threshold value of the quantized data other than the important position.

More preferably, the encoding means encodes the difference value between the previous frame and the current frame of the image data at the important position of the image, and encodes the original image of the current frame. It is characterized in that the boundary value for discriminating any of the above is determined so as to generate more (the mode for encoding the original image of the current frame) than the boundary value other than the priority position.

More preferably, the encoding means makes the generation period of the compulsory (mode for encoding the original image of the current frame) of the image data at the important position of the image shorter than the generation period other than the important position. It is characterized by

More preferably, the apparatus further comprises means for validating the already-registered priority range as it is even if the display of the priority position of the image is turned off.

It is possible to improve the transmission efficiency by an efficient encoding process while mainly improving the image quality of an image in an important range designated by the operator.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the arrangement of a TV telephone device according to an embodiment of the present invention. In the figure, 1 is a handset, which is one of the voice input / output means of this apparatus, and 2
Is a microphone, which is one of the voice input means of this device,
The speaker 4, which is one of the audio output means of the apparatus, has a function of switching the handset 1, the microphone 2 and the speaker 3 as the audio input / output means according to an instruction of the system control unit 12, and a gain adjustment function for adjusting the volume level. , An on / off detection function for detecting whether the handset 1 is in an on-hook state or an off-hook state, and a microphone 2 as a voice input / output unit.
The voice interface (i / f) unit 5 having an echo canceling function for canceling an echo when the speaker and the speaker 3 are used, and a tone generating function such as a dial tone, a ringing tone, a busy tone, and a ringing tone are designated by the system control unit 1.
64 kbps PCM (A-law),
64 kbps ADPCM (μ-law), 7 kHz audio (SB-ADPCM), 32 kbps PCM, 1
6 kbps (for example, APC-AB) has a function of A / D converting and encoding a transmission voice signal in accordance with a voice encoding / combining algorithm such as 8 kbps, and a function of performing D / A conversion of a reception voice signal in combination. It is a voice encoding / decoding unit.

Reference numeral 6 is a camera that is equipped as a standard equipment to image a person, 7 is one of the image inputting means of this apparatus, and a document camera for inputting pictures, drawings, etc. 8 is the camera 6 or the document camera A display unit for displaying an input image from the device 7, a received image from the other party, and an image from the system control unit 12;
Is a signal for electrically / signally matching the switching function of the image input means, the display switching and display division function of each image described above, and the video signal of each image input / output unit according to an instruction from the system control unit 12. Image input / output unit with conversion function,
Reference numeral 10 denotes an image encoding / decoding unit having a function of A / D converting and encoding a transmission image and a function of decoding and D / A converting a reception image. Motion compensation is performed on raw data of a large amount of image,
Frame drop, interframe compensation and interframe compensation, DCT
Bandwidth compression is performed by various methods such as conversion and vector quantization conversion to reduce the capacity and enable transmission over digital lines. At present, the basic interface of the ISDN line is 64 kbps, but CCITT Recommendation Draft H, 26 is available as an image encoding method that can be transmitted even at this transmission rate.
There is one.

Reference numeral 11 denotes an operation unit such as a keyboard or touch panel used for inputting control information for controlling the apparatus, and 12 denotes a CPU, ROM, RAM, auxiliary storage device,
A system control unit 13 that includes a character generator, an image signal generation circuit, and the like, monitors the state of each unit, controls the entire apparatus, creates an operation / display screen according to the state, executes an application program, and 13 is a voice encoding unit. Audio data from the decoding unit 5, image data from the video encoding / decoding unit 10, BA from the system control unit 12
SIT is a demultiplexing unit that multiplexes S in a transmission frame unit, separates a received frame into each medium of a constituent unit, and notifies each unit. CCITT recommendation is H.264. 221
There is.

Reference numeral 14 is a line interface for controlling the line according to the ISDN user network interface. Reference numeral 15 is a storage unit that records and stores the data from the audio i / f unit 4 and the image input / output unit 9 and can reproduce the stored data as necessary. 16 is a system control unit 12 This is an image synthesizing unit that synthesizes the image b from the image decoding unit, the image a from the image input unit, and the graphic image c from the graphic generating unit 17 according to the control signal d from S. Graphic generation unit 17
Generates graphic data from the CG and bitmap generated in the system control unit 12. 18
Is an emphasis range designation unit that stores the image emphasis position based on information from the system control unit 12 and instructs the graphic generation unit 17 and the image encoding unit to the emphasis position. 19 is a transmission screen to be transferred to the graphic generation unit 17. A storage unit for storing and managing the format and the like, the range of priority range designation and the processing method thereof, and the system control unit 12
It is provided inside.

Next, the operation will be described. When the operator selects a priority range designation mode from the operation unit 11, the system control unit 12 receives the selection mode and sets the first priority range to the priority range designation unit 18. Then, the priority range designation section 18 sends information to the graphic generation section 17 to indicate the priority range designation. The information is, for example, coordinate position information on the display screen of the emphasis range and information on the display method of the emphasis position.

The graphic generation unit 17 generates a display position of the graphic data and image information of the graphic data so as to be displayed on the screen based on the information, and transfers it to the image synthesis unit 16. The image information is not the color information that is actually displayed, but the index information of a table (color lookup table) in which color information that can be displayed simultaneously is selected from the displayable colors and registered.

Here, the internal structure of the image synthesizing unit 16 is shown in FIG.

Reference numeral 20 is a transmission video RAM for receiving and storing the transmission image from the image input unit, 21 is a reception video RAM for receiving and storing the reception image from the image decoding unit, and 22 is a transfer from the graphic generation unit 17. The color look-up table 23 for converting the index information into the color information of the image data is a transmission video RAM 20 and a reception video R.
AM 21 and a video sync generator for generating a vertical sync signal and a horizontal sync signal which are image data output timings from the color look-up table 22, a transmission video RA
The image data output from M20 is used as the system control unit 12
Image scaling unit that scales to a desired size according to instructions from
An image scaling unit 25 scales the image data output from the reception video RAM 21 to a desired size according to an instruction from the system control unit 12.

Reference numeral 26 is a color look-up for each pixel position (each display timing) on the screen of the display unit 8 based on the image data of the switching instruction of the video switch (SW) 27 instructed by the system control unit 12. Up table 22
The image data output from the transmission video RAM 20 is monitored, and the image data from the transmission video RAM 20 is monitored. When the image data is the image data that selects the image data from the transmission video RAM 20, the transmission video R is sent to the SW 27.
Control for selecting and outputting the image data from the AM 20. When the image data is image data for selecting the image data from the receiving video RAM 21, control for causing the SW 27 to select and outputting the image data from the receiving video RAM 21. In the case of other image data, the SW control unit, SW27, which controls to select and output the image data output from the color lookup table 22.
Selects one of the image data based on the SW control and outputs it to the image output unit.

The image synthesizing unit 16 displays the graphic data from the color lookup table 22 on the entire screen, and the window displays the transmission image data from the transmission video RAM 20 and the reception image data from the reception video RAM 21 on the screen. Can be made. Further, it is possible to display the graphic data from the color look-up table such that the transmitted image data is superposed on the received image data. Furthermore, when the transmission image data and the reception image data overlap each other, either of the images can be selected and displayed in the overlapping portion.

Next, FIG. 3 shows an internal configuration diagram of the image encoding unit in the image encoding / decoding unit 10.

Reference numeral 28 denotes an encoding control unit which performs discrimination processing according to an instruction from the system control unit in the overall control unit, controls each unit according to the discrimination result, and generates necessary information. Is a priority MB registration unit that takes in the priority position information from the priority range designating unit 18 and registers the MB of the emphasis position and the priority encoding control for the emphasis position, and 30 is an encoding target. A frame dropping control unit that suppresses the generated code amount by thinning out image frames, 31
Uses the same MB in the previous frame and the current frame to calculate the mean square error between frames and the variance within the frame, and then INT
ER (difference value between image of previous frame and current frame) / IN
INTER that determines TRA (the original image value of the current frame)
/ INTRA determination processing unit, 32 determines whether or not the target block of the current frame is searched within a certain range of the previous frame for pattern matching, that is, whether or not motion compensation is performed. Based on the determination result of the MC determination unit 32, the presence / absence of motion compensation and the motion vector (direction,
The MC control unit 34 that performs MC control by instructing the size) is an INTER / INTRA control unit that instructs INTER / INTRA control to each unit according to the determination result of the INTER / INTRA determination unit 31.

Reference numeral 35 is a filter control unit for instructing filter processing after motion compensation, and reference numeral 36 is a decoded image due to a quantization error accumulated on the receiving side when INTER continues for the same MB, or a transmission error. Since a problem such as irrecoverability cannot be obtained when a disturbance occurs, a refresh cycle counter 37 for refreshing by setting to INTRA in a certain period and 37 is a threshold for improving the generation efficiency of data when quantized. A quantizing threshold value control unit for controlling the value and setting the value of data equal to or less than a certain value as 0 value so as not to generate the data, 38 is a transmission buffer 49.
Quantization step size control unit for controlling the quantization step size in accordance with the data storage amount of, a transmission amount detection unit for detecting the storage amount of the transmission buffer 49, 40 for a frame header, a GOB header, and an MB header. The header information generation unit to be generated, 41 is a frame memory, and 42 and 43 are
A switch for selecting INTER / INTRA, 44 is
It is a subtracter that takes the difference between the previous frame and the current frame at the time of INTER.

Reference numeral 45 denotes a DCT transform unit (lossy coding) for transforming from the spatial domain to the frequency domain by orthogonal transform, 46.
Is a quantizer, 47 is a VLC unit (reversible coding) for variable-length coding the generated data, 48 is a multiplexing unit, 49
Is a transmission buffer, 50 is a BCH unit that generates an error correction frame, 51 is an inverse quantizer, 52 is an inverse DCT transform unit, 53 is an adder that adds the previous frame and the current frame, 54, Reference numeral 55 is a switch for alternately switching between a read frame and a write frame, 56 and 57 are frame memories, 58 is an FM control unit for controlling the switches 54 and 55 and the frame memories 56 and 57, and 59 is a search for the previous frame. Then, the motion compensating units 60 and 61 for detecting the pattern matching with the current frame are ON / OF of the filter.
A switch for selecting F, 62 is a filter processing unit, 63
Turns ON / OFF the data input to the FIFO (memory) 64.
This is a switch that turns off.

Regarding the structure of the image to be encoded,
As shown in FIG.

H. In the H.261 Recommendation Draft, a plurality of different standards such as NTSC, PAL, and digital television standards exist for the video signal to be handled, so that a video signal format common to the world is adopted so that they can communicate with each other.

This format is called the CIF format, and the number of samples is 352 pixels × 288 lines with luminance Y.
The color differences Cr and Cb are defined by 176 pixels × 144 lines.

Regarding the sample points (sampling points), the color differences (Cr, Cb) are the luminance 4 points (Y1, Y2, Y3).
Y4) are defined as equidistant points.

Further, 1/4 of CIF is called QCIF format, and the number of samples is defined as 176 pixels × 144 lines for luminance Y and 88 pixels × 72 lines for color differences Cr and Cb.

Here, the above-mentioned format is composed of a plurality of GOB formats, and the GOB format is composed of 33 MB formats. Further, in the MB format, the luminance blocks of 8 pixels × 8 lines are Y1, Y.
The block is composed of 4 blocks of 2, Y3 and Y4 and the color difference block of 8 pixels × 8 lines and 2 blocks of Cr and Cb, and has a hierarchical structure.

With this hierarchical structure, encoding can be performed in MB units.

The GOB has a sample size of luminance 176 pixels × 48.
Line, color difference Cr, Cb 88 pixels × 24 lines, defined as 1/12 of CIF and 1/3 of QCIF, and G
OB number is CIF, GOB1 to GOB12, Q
GOB1, GOB3, and GOB5 are assigned to the CIF.

Next, the frame structure of the encoded image data has a multiplexed frame structure as shown in FIG. However, for convenience of description, the description will be made with FH added.

FIG. 5A shows the structure of GOB blocks. In this way, FH is attached to the beginning of the data of one frame of the screen, and the block obtained by dividing the screen into 12 is G
GOB1 to GOB12 are sequentially transmitted as OB.

The GOB division method is shown in FIG. FIG. 5B shows the details of the FH and GOB shown in FIG. FH is composed of PSC, TR, and PTYPE. The PSC is a frame start code and has 20 bits of "0000 0000 0000 0001 00".
00 ". TR is a frame number of 5 bits" 1 ".
Values from 0 to “30” are used. PTYPE is 6-bit type information, split screen instruction information,
The document camera instruction information, screen freeze release, and information source format instruction information (CIF, QCIF) are included. That is, the FH decoding unit notifies the control unit 28 of the result of decoding the above contents.

Next, the GOB header is composed of GBSC, GN and GQUANT. GBSC is a GOB start code of 16 bits "0000 0000 0000".
0001 ". The GN is a GOB number, which is 4 bits and is used from" 1 "to" 12 ". When the GN is" 0 ", it is used by the PSC, so the FH PSC and GOB are used.
GBSC + GN can be regarded as a continuous value of 20 bits. GQUANT is 5 as the quantization characteristic information.
Includes quantization step size information in bits.

The MB header is a macroblock (hereinafter referred to as MB
(Referred to as)) is a header of a pixel block.

As described above, MB is 1 in 33 MBs.
One GOB is composed of four luminance signals (Y) of 8 pixels × 8 lines and two color difference signals (Cb, Cr) of 8 pixels × 8 lines. Here, the number of each block is 1 to 4 for Y, 5 for Cb,
The number 6 is assigned to Cr. This MB header has MBA, MTYPE, MQUANT, MVD and CBP.
It is composed of.

MBA is a macroblock address representing the position of the MB, and only the first MB is an absolute value, and thereafter, the variable length code of the difference. MTYPE is the type information of the MB, and the data of the MB such as INTRA (intra-frame encoding), INTER (inter-frame differential encoding), MC (motion compensation with inter-frame differential encoding), FIL (filter) is processed. The specified processing type has been inserted. MQ
UANT is the same as GQUANT. MVD is motion vector information. The CBP is a significant block pattern and includes, as information, the four Ys of MB and the number of an effective pixel block of Cr and Cb. After the MB header, as described above with respect to the compression-coded image data, the pixel blocks that are significant blocks among the Y4 Cr and Cb are arranged in the numerical order.

Next, the data output to the line is an error correction frame having a format as shown in FIG. 1
The frame is composed of error correction frame bits of 1 bit, fill identifier of 1 bit, pixel data of 492 bits, and error correction parity of 18 bits, that is, 512 bits in total. Further, this frame is composed of 8 frames to form one multi-frame.

This image compression method is based on H.264 standard in CCITT.
It has been made into a recommendation as the H.261 recommendation, and if it complies with the recommendation, it is possible to perform mutual communication with a TV phone that complies with other recommendations.

The encoding method will be briefly described. First, the data in the frame is divided into 8 by utilizing the fact that the correlation between pixels is strong in the image in the natural world and the frequency components are concentrated in the low frequency and the high frequency is small. Intra-frame coding that performs a two-dimensional DCT conversion as a block of 8 pixels and when there is a strong correlation between the two in the image block at the same position of the previous frame and the current frame, the difference between the frames is calculated and 8 pixels are added to the difference value. Inter-frame coding in which a block of × 8 pixels is two-dimensionally DCT-transformed, and when a similar image block relatively moves from the previous frame to the current frame, this is detected and the movement amount and movement of the image block are detected. Motion compensation that reduces the amount of data generated by sending only the direction information and not the image data itself, and the coefficient value for each frequency after DCT conversion is low frequency. A value is generated in the wave domain, but it is difficult to generate a value in the high frequency domain.Zero run length coding that utilizes the fact that zero value continues and by changing the quantization step width of data according to the amount of data generated Quantization that adjusts the amount of data generated and by assigning short code values to data patterns that occur frequently and long code values to data patterns that occur less frequently Variable length coding to convert to a smaller amount of data and multiple compression techniques such as frame dropping that skips frames and drops the image data itself are hybridized so that moving images can be communicated even at low rate communication There is.

Next, the method of designating the priority range is as follows.
As shown in FIG. 7, the priority range is designated in MB units. Figure 7
The area indicated by the coarse dots indicates the second priority, and the area indicated by the diagonal lines indicates the first priority.

A frame surrounding this area is transferred from the emphasis range designating unit 18 of FIG. 1 to the graphic generating unit 17, the graphic generating unit 17 transfers the frame information to the image synthesizing unit 16, and the image synthesizing unit 16 The frame information is combined with the self-portrait of the video signal a from the camera, transferred to the image output unit, and displayed on the display unit 8 from the image output unit to notify the operator of the important range. A concrete image of the display screen is shown in FIG.

In FIG. 8A, the partner image is displayed as a full screen and the self image is displayed as a sub screen, and the self image is enlarged in FIG. 8B. In (b), a frame indicating the emphasis range is displayed. The area within this frame shows the priority range.

(C) is an example in the case where a plurality of emphasis positions are designated. As described above, when a plurality of subjects are the targets, the coding efficiency can be improved by designating a plurality of priority ranges and performing the priority coding. (D) is
This is the case where the priority range is specified at multiple levels. By designating the inner frame as a higher priority level than the outer frame, it is possible to perform more precise and more intensive encoding.

FIG. 9 shows an MB control management table. This MB control management table is a table that manages how to control encoding for each MB.

The MB No. designation method is GOB No. and STAR.
It is specified and managed by T MB No. and END MB No. The START BNo. Is the MBNo. Where the priority position within the GOBNo. Started, and the END MBNo. Is the MBNo. Where the priority position within the GOBNo. This START MB No.
The MB between the END MB No. and the END MB No. is also the MB in the priority range. Therefore, when there are multiple range designations and multiple level designations, multiple START MBNo.s within the same GOBNo.
And END MB No. are specified.

The MB thus determined defines the INTER / INTRA discrimination standard, the quantization threshold standard, the quantization step size standard, and the refresh cycle standard according to the priority level thereof, and is controlled by each control standard. An MB for which the priority level is not applicable is determined to be a normal MB and controlled.

The management table of FIG. 9 is an example created based on the priority position designation of FIG. The INTER / INTRA discrimination standard is shown in FIG. FIG. 10A shows the INTER / INTRA discrimination standard in the normal case. The vertical axis is the variance value of the difference within the block (B) of the current frame of the target MB, which is defined by the following equation.

[0066]

F-dif = ΣB ² / 256- (ΣB
/ 256) ^{2 The} horizontal axis is the mean square error value of the difference between the motion-compensated block (MCB) of the previous frame of the target MB and the block (B) of the current frame, and is defined by the following equation. .

[0067]

[Number 2] mc-dif = Σ | B- MCB | as ^2/256 discrimination criterion, if less than a certain value mc-dif unconditionally is determined INTER mode, f-
When dif ≧ mc-dif, INTER, f-dif
When <mc-dif, it is determined to be INTRA.

In (b), the value of mc-dif for unconditionally discriminating the INTER mode is set to a value smaller than usual so that f-dif ≧ m even when mc-dif is small.
When c-dif, INTER, f-dif <mc-d
When it is “if”, it is a discrimination criterion when the probability of occurrence of INTRA is higher than usual by discriminating from INTRA. This is because INTR is used when the difference in the frame is small.
By setting A, it is possible to eliminate the quantization error on the receiving side in a state where the amount of generated data is small.

In (c), the slope of the boundary line between INTER and INTRA is made larger than usual, so that in the region surrounded by the boundary between the dotted line and the solid line, that is, more than the normal case, the ratio of the increase in slope is larger than INTRA. This is the INTER / INTRA discrimination standard when it is discriminated. Therefore, the above proportion is discriminated as INTRA, and the quantization error on the receiving side can be eliminated by the occurrence probability, and the image quality can be improved. In particular, since the probability of occurrence of INTRA increases as mc-dif increases, it is possible to reduce the accumulation of quantization errors when the difference value is large.

(D) shows INTER / IN
By shifting the boundary line of TRA in parallel to the smaller mc-dif, in the region where the boundary line is displaced,
INTER / I when INTRA is generated
This is the NTRA discrimination standard. That is, INTER and INT
Since the threshold value of the boundary at which RA occurs is shifted by a fixed value toward the higher f-dif, only the shifted value is:
Since INTRA occurs, the probability of being reliably discriminated as INTRA increases regardless of the value of mc-dif.

As described above, by causing INTRA to occur more than usual, it is possible to suppress the accumulation of quantization errors on the receiving side as much as possible, and it is possible to refresh image quality deterioration due to decoding errors and the like as soon as possible. Become.

Since the quantization threshold standard of FIG. 9 adopts the method of suppressing the data generation amount by performing the zero run length coding on the coefficient value which becomes 0 after the quantization.
The maximum value of the variable threshold value control for generating the coefficient of 0 as continuously as possible is shown. However, since it is a scaling factor for the set quantization step size, the value obtained by multiplying the quantization step size by this scaling factor becomes the maximum value of the quantization threshold value. Usually, the coefficient is doubled and the threshold value is increased to generate more coefficient 0 values. Therefore,
By reducing the quantization threshold, the coefficient value that was normally rounded to 0 will also have a value, and more detailed coefficients can be sent, so the reproducibility of the image at decoding Will improve.

Next, with reference to the quantization step size shown in FIG. 9, normally, when no special image block such as an emphasized image exists, each image block is the same as shown in FIG. 11A. Quantization processing is performed.

The vertical axis of FIG. 11 represents the quantization step size. The value varies from 2 to 62.
The horizontal axis represents the degree of sufficiency of the transmission buffer. The buffer fullness is defined as the amount of data generated in a frame (in H.261 recommendation, in QCIF, 64 kbit).
In CIF, it is defined as 256 kbit),
It is a value obtained by accumulating the increase and decrease of the difference from the actually generated data amount based on the expected generated data amount when averagely allocated to each block based on the data amount.

(A) shows the case where the quantization step size changes linearly with respect to the degree of sufficiency. That is, as the amount of generated data increases and the degree of sufficiency increases, the quantization step size is increased to suppress the amount of generated data. On the contrary, as the amount of generated data is small and the degree of sufficiency is small, the quantization step size is reduced and the image quality is improved.

In (b), the quantization step size processing of the normal image block and the image block at the priority position is divided, and in the same transmission buffer sufficiency, in the normal case, the quantization step size is increased to increase the priority. In case of position,
By reducing the quantization step size, more data is assigned to the image block at the important position.

(C) is a case where there are a plurality of priority levels,
In this case, the priority level 1 is higher than the priority level 2. In the normal case, the quantization step size reaches the maximum value before the transmission buffer fullness is full, so that the image quality is considered to be considerably rough.

(D) is an example in the case where the image quality at the normal position and the image quality at the emphasis position are made extremely extreme. Since there is a considerable difference compared to (c), it seems that the image quality at the important position is improved compared to (c). In the example of FIG. 9, (c) of FIG. 11 is used.

Next, referring to the refresh cycle reference of FIG. 9, since it is not possible to improve the transmission of only the difference value in the INTER mode due to a quantization error on the receiving side, an error in the received data, etc., normally up to 132 times. The original image (INTRA) is transmitted to the other terminal to refresh the image and respond. Therefore, the priority of the refresh cycle is set shorter in the priority position than in the normal position (that is, the maximum number of times of refreshing is reduced), thereby providing superiority and inferiority. In the example, the priority level 1 is 66 times, and the priority level 2 is 110 times.

By referring to the above-described processing of FIG. 9 every time each macro block is processed, the macro block at the priority position is subjected to the processing of improving the image quality as compared with the normal macro block. By doing so, it becomes possible for the operator to specify the emphasis position while looking at the monitor, and automatically execute the control for the specified emphasis position with the image quality improved compared to the normal position.

Although the TV telephone apparatus has been described above as an example, the present invention is not limited to this, and can be similarly applied to any apparatus for communicating moving images.

[0082]

As described above, according to the present invention,
While looking at the display screen, it is possible to improve the image quality of the image at the important position of the image transmitted to the other terminal.

Therefore, since the coded data of the image can be preferentially assigned to the image area designated by the operator in the priority range, accurate and appropriate coding control is performed to improve the image quality. This has the effect of dramatically improving the coding efficiency. Further, there is also an effect that the convenience of the operator is improved because the image quality can be surely improved while visually checking according to the operator's request.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is an internal configuration diagram of an image composition unit.

FIG. 3 is an internal configuration diagram of an image encoding unit.

FIG. 4 is a diagram showing an image format.

FIG. 5 is a diagram showing an image data multiplexed frame.

FIG. 6 is a diagram showing error correction frame synchronization.

FIG. 7 is a diagram showing priority position designation.

FIG. 8 is a diagram showing an example of a priority image designation display.

FIG. 9 is a diagram showing an MB management table.

FIG. 10 is a diagram illustrating INTER / INTRA discrimination control.

FIG. 11 is a diagram illustrating quantization step size control.

[Explanation of symbols]

 1 Handset 2 Microphone 3 Speaker 4 Voice i / f unit 5 Voice coding / decoding unit 6 Camera 7 Document camera 8 Display unit 9 Image input / output unit 10 Image coding / decoding unit 11 Operation unit 12 System control unit 13 Demultiplexing Section 14 line i / f section 15 storage section 16 image composition section 17 graphic generation section 18 priority range designation section 19 storage section

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location H04N 5/66 B 7/08 7/081

Claims (9)

[Claims] 1. A display means for displaying a transmission image, a setting means for displaying information indicating a focus position of the image on the displayed transmission image and interactively setting the focus position, and a focus position of the image. An image transmitting apparatus comprising: an encoding unit that performs encoding for improving the image quality of the image data inside the image data compared to the image data at other positions. 2. The image transmitting device according to claim 1, further comprising a designating unit that designates at least one focus position of the image. 3. The image transmitting apparatus according to claim 2, wherein the designation unit designates the priority position of the image at a plurality of levels. 4. The image transmitting apparatus according to claim 2, wherein the designation unit designates the emphasis position of the image in units of minimum image blocks for encoding. 5. The image transmitting apparatus according to claim 1, wherein the encoding unit makes the quantization step width of the image data at the important position of the image smaller than the quantization step width at a position other than the important position. 6. The encoding device according to claim 1, wherein the encoding unit makes a threshold value of quantized data of image data at an important position of the image smaller than a threshold value of quantized data at a position other than the important position. Image transmission device. 7. The encoding device according to claim 1, wherein the encoding unit encodes (a mode for encoding a difference value between a previous frame and a current frame) of image data at an important position of the image and (an original image of the current frame. Mode)
The image transmitting apparatus is characterized in that the boundary value for discriminating any of the above is determined so as to generate more (mode for encoding the original image of the current frame) than the boundary value other than the priority position. 8. The encoding device according to claim 1, wherein the encoding unit generates a compulsory generation mode (a mode for encoding an original image of a current frame) of image data at a priority position of the image other than the priority position. An image transmitting device characterized by being shorter than the cycle. 9. The image transmitting apparatus according to claim 1, further comprising means for validating the already-registered priority range as it is even if the display of the priority position of the image is turned off.