JP3197894B2 - Image transmission device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[Industrial applications]

 The present invention relates to an image transmission device such as a videophone device.

[0002]

[Prior art]

Still video telephones are black and white still video telephones (see Video Information (I) 1988/9) that conform to the unified standard (JJ-40.10) established by the Telegraph and Telephone Technical Committee (TTC). The main specifications are: connection line: analog telephone line Modulation method (modem): amplitude and phase modulation (AM-PM) Transmission pixel size: mode A (100 (V) x 160 (H),
Mode B (96 (V) × 100 (H)) Transmission time: Mode A (10 seconds), Mode B (about 6 seconds). In addition, semi-video video phone equipment compatible with ISDN (Digital Communication Network) (see Video Information (I) 1988/9) is 64kbps
Is a simple video-telephone device that uses a digital line. Simple video (semi-video) means that a low-resolution image of about 128 × 120 pixels can be moved in a so-called “frame-dropping” manner at about 10 frames / sec. Means that the amount of transmitted information is significantly reduced. In this semi-movie video telephone device, the image distortion is very conspicuous, but the movement can be recognized temporarily.

[0003]

[Problems to be solved by the invention]

The above-mentioned TTC-compliant black-and-white still image video telephone device transmits a black-and-white still image taken by a video camera attached to the main body through a general analog telephone line, but the image that can be transmitted is 100 × 160 or 96 × pixels. It is a low resolution of 100, the image transmission time is 6 to 10 seconds, and the voice call is interrupted during image transmission. Since the still image transmission is performed, the time during which the voice call is interrupted is a complete communication waste time. It is better to be as short as possible.

[0004] Since the high-speed digital line can be used in the above-described ISDN-compatible semi-video video phone, simultaneous transmission of voice and semi-video is possible by using voice compression technology. However, in order to achieve video transmission, we make full use of image data compression technologies such as motion compensation, intra- and inter-frame adaptive prediction coding, vector quantization, and conditional pixel interpolation, and are about 128 x 100 pixels at 10 frames / sec. Since the amount of transmission needs to be compressed before, when the data amount of the original image is small, the compression rate can be spared and the image quality does not deteriorate much, but when the data amount of the original image is large, The image quality deteriorates significantly.

An object of the present invention is to provide an image transmission apparatus which can improve the above-mentioned drawbacks and reduce the amount of transmission information without deteriorating the image quality of a portion of interest on a screen.

[0006]

[Means for Solving the Problems]

In order to achieve the above object, the invention according to claim I is based on image input means for inputting an image, area recognition means for recognizing an area to be transmitted in one screen of the image input by the image input means, A transmitting unit for transmitting an image of the region recognized by the region recognizing unit, wherein the region recognizing unit converts the image input by the image input unit into blocks each having a predetermined number of pixels. The output block is divided and subjected to orthogonal transformation to obtain an output block. The output block is used to calculate an evaluation value relating to the distribution of high frequency components of the output block using a weight coefficient block. The block 4 which is equal to or larger than the value is determined as an area of the one screen to be transmitted.

[0007]

[Action]

In the invention described in claim I, the image input means inputs an image, the area recognition means recognizes an area to be transmitted in one screen of the image input by the image input means, and the transmission means recognizes the area by the area recognition means. The image of the specified area is transmitted. In this case, the region recognizing unit divides the image input by the image input unit into blocks each having a predetermined number of pixels, performs an orthogonal transform on each block to obtain an output block, and obtains a weight coefficient block for the output block. Is used to calculate an evaluation value relating to the distribution of the high frequency component of the output block, and a block whose evaluation value is equal to or greater than a predetermined value is determined as an area to be transmitted in the one screen.

[0008]

【Example】

 FIG. 1 shows an embodiment of the present invention.

This embodiment is an example of a videophone device, and a video camera 21 constituting an image input unit captures an image of a subject and converts the image into an analog image signal 22. The video signal input circuit 23 converts the analog image signal 22 from the video camera 21 into a digital image signal 24 by sampling and performing analog / digital conversion. The image memory 25 has a three-port configuration, and performs writing and reading of the digital image signal 24 from the video signal input circuit 23, and writing and reading of data through the system bus 26.
Further, the image memory 25 has a function of a display frame memory, and data in the image memory 25 is read out as a digital image signal 27 and input to the video signal output circuit 28. The video signal output circuit 28 converts the digital image signal 27 from the image memory 25 into an analog video signal 29 and outputs it to the display 30. The display 30 displays an image using the analog video signal 29.

The line control circuit 31 includes an analog telephone line, a telephone, and a modem.
Control 32 connections. Normally, when making a voice call using a telephone, the telephone and the analog telephone line are connected to a line control circuit.
In the case of performing image transmission, the modem 32 and an analog telephone line are connected by a line control circuit 31. The modem 32 modulates the data sent from the system bus 26 into a voice band signal and sends it to the analog telephone line via the line control circuit 31, and conversely, the data sent from the analog telephone line via the line control circuit 31 The signal is demodulated and output to the system bus 26. The discrete cosine transform calculator 33 performs a discrete cosine transform (DCT) of the image data and its inverse transform.
An operation switch 34 gives an operation instruction of the present embodiment by a caller, and a system controller 35 performs operation control of the entire apparatus, operation of the image memory 25, image data processing, and the like.

A video signal 22 captured by a video camera 21 is input to a video signal input circuit 23, converted into a digital image signal 24, and temporarily stored in an image memory 25 in real time. The digital image signal stored in the image memory 25 is immediately read out, converted into an analog video signal 29 by the video signal output circuit 28, and sent to the display 30 for image display. Since these operations are performed in real time, a moving image monitor is possible.

When the caller instructs image transmission by the operation switch 34, the system controller 35 detects the instruction by a signal from the operation switch 34, and first writes the digital image signal 24 from the video signal input circuit 23 to the image memory 25. (This is called an image freeze). Next, the system controller 35 compresses the frozen still image (a digital image signal for one screen stored in the image memory 25), sends the data to the modem 32, and converts it into an audio band signal. The voice band signal is sent to the analog telephone line through the line control circuit 31 and transmitted to the destination. At this time, if the purpose is to transmit the face or bust of a person, it is not necessary to transmit the entire screen of the frozen still image, and only the face or bust of the frozen still image is extracted. And send it. In fact, in a videophone device, the area of a person image to be transmitted is generally a very small area as compared with the entire screen, and if only the image of the area is extracted and transmitted, the transmission data amount is the entire screen image. Can be significantly removed compared to the case. That is, in the still picture videophone device, the transmission time is shortened by the reduced amount of data, the time during which the voice call is interrupted is shortened, and the communication waste time is reduced. Therefore, in this embodiment, only a part of the frozen still image is extracted and transmitted.

Next, a method of extracting only a part of a still image in this embodiment will be described.

In a videophone device or the like, the area to be transmitted on the screen is the focused area of the screen, and when capturing an image with the video camera 21, the focused area should be sufficiently focused. The part is in a so-called blurred state as compared with the attention part. Therefore, if a focused portion of the screen is detected, a portion of interest can be extracted. As a method of detecting a focused portion of a screen, there is a method of detecting a high-frequency portion of image data. The discrete cosine transform operation unit 33 can perform the following two-dimensional DCT and its inverse transform.

[0015]

(Equation 1)

By performing the DCT on a certain pixel block, the distribution of a high frequency portion can be known. It is assumed that the output pixel block S is obtained by performing the DCT on the input pixel block P as shown in FIG. However,
FIG. 2 shows an example of a 4 × 4 block (N = 4), where Pi, j =
x (i-1, j-1) and Si, j = x (i-1, j-1).
The components Si, j (i, j = 1 to 4) of the output pixel block S are i =
When j = 1, represents the DC component of the input pixel block P, and i,
Higher frequency components are represented as j increases. Therefore, the second
A weight coefficient block W (the component is Wij) as shown in the figure is determined, and the distribution of the high-frequency portion + minute is evaluated by the following equation.

[0017]

(Equation 2)

Then, as shown in FIG. 2, v is compared with a constant value f _TH by the comparing means 20, and if v is a constant value f _TH, it is determined to be a focused portion. That is, it is determined that v ≧ f _TH : focused portion v <f _TH : blurred portion. By transmitting only the portion of the screen where v ≧ f _TH , only the target portion of the screen can be transmitted. In this case, the attention area of the screen often occupies a certain continuous area in the entire screen such as the face of a person, and is continuous by a certain number of input pixel blocks determined to be in focus. If a block to be transmitted is determined, a valid target block can be detected more reliably.

In this embodiment, when data of a frozen still image is compressed and sent to the modem 32, the system controller 35 reads out a pixel block from the image memory 25 and sends it to the discrete cosine transform calculator 33 as shown in FIG. Input pixel block P
And let the DCT be performed as described above. Next, the system controller 35 outputs the DC from the discrete cosine transform
The pixel block subjected to T is read as an output pixel block S, and the calculation with the weight coefficient block W is performed as described above to calculate v. This v is compared with a constant value f _TH to obtain a constant value f. If it is equal to or greater than _TH, it is determined that the block is to be transmitted, and the output pixel block is set as encoded data that is data compressed.

FIG. 4 shows an example of a freeze screen image stored in the image memory 25. In the figure, the shaded blocks are the portions determined as the attention areas by the above-mentioned method, and the unshaded blocks are other areas (areas not to be transmitted). If one row in the scanning line direction (horizontal direction) of one pixel block is called a block row, in the example of FIG. 4, the first block row and the second block row are all parts that do not transmit, and the third block Below the row, pixel blocks not to be transmitted continue for a plurality of pixel blocks, pixel blocks to be transmitted next continue, and pixel blocks not to be transmitted again continue for a plurality of pixel blocks, and the block row ends. As described above, in this embodiment, the image should be divided into blocks each having a predetermined number of pixels, the block should be evaluated to determine the region of interest, and the block including the outline of the region of interest and the blocks inside it should be transmitted. Extract as a region.
In this embodiment, the transmission format is determined on a block row basis as described above. The transmission format is shown in FIG. The frame unit is one block row unit. The transmission format is as shown in FIG. 5A when there is a pixel block to be transmitted, and as shown in FIG. 5B when not transmitting all the block rows. However, if the number of transmission pixel blocks can be determined when the coded image data portion is restored, the number of blank blocks (2) can be omitted. As shown in FIG. 6, the system controller 35, following the frame identification flag at the beginning of the frame, follows the block line number, the number of the first consecutive non-transmitted pixel blocks (the number of blank blocks (1)), and the code of the pixel block to be transmitted. The sequence is sent to the modem 32 in the order of the frame image sequence, the number of pixel blocks not to be transmitted (the number of blank blocks (2)), and the frame check sequence (FCS).

Although the transmission format in FIG. 5 shows an example in which a pixel block to be transmitted appears only once for one pixel block, a case where a pixel block to be transmitted appears twice or more for one pixel block is also illustrated. It is determined similarly.

In the case of reception, a voice band signal transmitted by an analog telephone line is input to a modem 32 through a line control circuit 31, where it is demodulated and sequentially converted into data of a transmission format shown in FIG. Output to 26.
This data is restored on the image memory 25 as a received image by the reverse operation of the above-described transmission, and is displayed on the display 30. That is, when the system controller 35 detects the first flag of the frame from the data output from the modem 32 to the system bus 26, it knows the next block line number. The received data from the modem 32 is sequentially written. In this case, the system controller 35 first writes white data to the block rows on the image memory 25 by the number of blank blocks (1). Next, the system controller 35 converts the coded image data from the modem 32 into the original pixel block by performing the inverse transform of DCT with the discrete cosine transform calculator 33 for each pixel block, and sequentially restores the block rows on the image memory 25. Write to. When this restoration is completed, white data is written into the block rows on the image memory 25 by the number of the next blank blocks (2), and the reception of the block rows is completed.

[0023]

【The invention's effect】

As described above, according to the first aspect of the present invention, image input means for inputting an image, area recognition means for recognizing an area to be transmitted in one screen of the image input by the image input means, A transmitting unit for transmitting an image of the area recognized by the area recognizing means, wherein the area recognizing means divides the image input by the image input means into blocks each having a predetermined number of pixels. Then, each block is subjected to an orthogonal transformation to obtain an output block, and an evaluation value relating to the distribution of the high-frequency component of the output block is calculated for the output block by using a weighting coefficient block. Since the above block is determined as the area to be transmitted in the one screen, the image quality of the noticed part to be transmitted in one screen of the image input by the image input means is degraded. The amount of transmitted information can be reduced, and only the target portion to be transmitted can be extracted from one screen of the image input by the image input means, so that the amount of transmitted information can be further reduced. The attention portion to be transmitted can be automatically extracted from one screen of the image, and the operability is greatly improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a conceptual diagram showing a high-frequency component detection method of the embodiment, FIG. 3 is a flowchart showing a processing flow of a system controller in the embodiment. FIG. 4 is a view for explaining image transmission in the embodiment, FIGS. 5A and 5B are diagrams showing a transmission format in the embodiment, and FIG. 6 is a processing flow of a system controller in the embodiment. It is a flowchart which shows.

[Explanation of symbols]

21 Video camera 23 Video signal input circuit 25
... Image memory, 28 ... Video signal output circuit, 30 ... Display unit, 31 ... Line control circuit, 32 ... Modem, 33 ... Discrete cosine transform operation unit, 34 ... Operation switch, 35 ... System controller , P ... input pixel block, S ... output pixel block, W ... weight coefficient block, 20 ... comparison means.

Continuation of the front page (58) Field surveyed (Int. Cl. ⁷ , DB name) H04N 1/38-1/393 H04N 1/32

Claims (1)

(57) [Claims] An image input means for inputting an image, an area recognizing means for recognizing an area to be transmitted on one screen of the image input by the image input means, and an area recognizing means for recognizing the area recognized by the area recognizing means. An image transmitting apparatus comprising: a transmitting unit that transmits an image, wherein the area recognizing unit divides the image input by the image input unit into blocks each having a predetermined number of pixels, and performs orthogonal transform on each block. To obtain an output block, calculate an evaluation value relating to the distribution of the high frequency component of the output block using the weight coefficient block for the output block, and assign a block having the evaluation value equal to or greater than a predetermined value to the one screen. An image transmission apparatus characterized in that it is determined as an area to be transmitted.