JPH0472984A - Moving picture transmitter - Google Patents

Abstract

PURPOSE:To obtain a picture easy to see automatically even from a video image picked up under a band lighting condition by detecting a moving region so as to apply brightness conversion depending on the brightness distribution of the above moving region. CONSTITUTION:A moving region detection circuit 101 has a frame memory and stores a reference picture data such as a preceding frame data. The circuit 101 discriminates a moving region and a still region by comparing the inputted frame data with the reference picture data and gives the result of discrimination to a separation circuit 102. The moving region signal is inputted to a signal distribution calculation circuit 103, in which the signal distribution of the moving region is calculated. A brightness conversion circuit 105 decides the conversion characteristic based on the received calculation result to apply brightness conversion to the moving region picture data.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a moving image transmission device used for transmitting moving images in video conferences, video telephones, and the like.

(Prior Art) With the spread of digital lines, the development and popularization of video communication equipment such as video conferences and video telephones is gaining momentum.

In such devices, lighting is also an important point in improving image quality. For example, in the case of fixed equipment such as video conference equipment, various lighting conditions must be adjusted, but it is difficult to provide appropriate lighting to the entire screen. The problem of lighting is even more serious for something like a videophone, where you don't know where it will be placed. In the case of a home, it would be difficult to expect users to use their own dedicated lighting, and there is a possibility that they will be used while moving. To address this problem, for example, currently commercially available still image videophones allow the user to adjust the sensitivity of the input camera using a knob.

However, with conventional video image transmission devices, 1) users have to adjust the screen each time they make a call, which is cumbersome; For example, there is no appropriate adjustment position for situations such as (i.e., when it is dark, there is no contrast on the face, and when it is bright, the background light is too bright and dazzling), so it cannot be dealt with. There was a problem that sometimes people appeared dark.

(Problem to be Solved by the Invention) As mentioned above, in the conventional moving image transmission device, the user must adjust the brightness of the screen in order to obtain an easy-to-see image, and this adjustment does not affect the entire screen. Therefore, there was a problem in that appropriate adjustment could not be made especially at the time of a reverse sale.

The present invention has been made to solve the above-mentioned problems, and aims to provide a moving image transmission device that can automatically obtain easy-to-see images even when images are taken under poor lighting conditions.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above-mentioned object, the moving image transmission device of the present invention comprises: a moving area detecting means for detecting a moving area or a skin area from an input image; The apparatus includes a brightness distribution calculating means for calculating the brightness distribution in the moving area by the moving area detecting means, and a brightness converting means for converting the brightness and chromaticity according to the calculation result by the brightness distribution calculating means.

(Function) According to the moving image transmission device of the present invention configured as described above,
By detecting a moving area and performing brightness conversion according to the brightness distribution of the moving area, the contrast of the moving area can be increased and the screen can be easily viewed.

In addition, by converting the brightness of the moving area and other areas, that is, the static area, separately, it is possible to detect cases such as backlighting, where the static area is much brighter than the moving area. Processing such as reducing the brightness of the background can be performed to make the screen easier to see.

Furthermore, by converting the color tone of the screen so that the skin area has the correct skin color, it is possible to get close to the correct color tone even when using the device in a dimly lit area.

(Example) Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the overall configuration of a moving image transmission device according to the present invention. The input image signal in the figure is input to, for example, a CIF and QCIF moving area detection circuit 101 and a separation circuit 102 in the standard system. The moving area detection circuit lot has a frame memory therein, and stores reference image data such as previous frame data. The input frame data is compared with this reference image data to determine whether it is a moving area or a static area, and the determination result is sent to the separation circuit 10.
Send to 2. This determination result may be an on10ff signal for each pixel synchronized with the image input, or if the determination result is for each block, the on/o f f for each block.
It is also possible to output each frame in the form of a 0/1 map. The separation circuit 102 separates the input image into a moving region and a still region according to this determination result, and supplies the separated regions to subsequent circuits separately.

The subsequent circuitry includes a signal distribution calculation circuit 103, 104, a luminance conversion circuit 105, and a 10B synthesis circuit 107. First, the dynamic area signal is input to the signal distribution calculation circuit 103, where the signal distribution of the dynamic area is calculated. This signal distribution is
It may be something like luminance histogram, pixel value dispersion, or P-P difference. After the calculation, the signal distribution calculation circuit 103 outputs the calculation result and the moving area image data to the brightness conversion circuits 105 and 106. Brightness conversion circuit 10
5 determines conversion characteristics from the received calculation results and converts the brightness of the moving area image data. Similarly, the stationary area signal is also sent to the signal distribution calculation circuit 104 and the luminance signal conversion circuit 10B for processing. Since the calculation results of the signal distribution in each area are used to determine the brightness conversion characteristics in both areas, the calculation results in the moving area are also sent to the brightness conversion circuit 10G, and the calculation results in the static area are also sent to the brightness conversion circuit 105. It will also be sent to

The luminance-converted moving region image data and still region image data are recombined into one original frame in a combining circuit 107, encoded in an encoding circuit 108, and transmitted.

Next, FIG. 2 shows an embodiment of the dynamic area detection circuit 101 of FIG. 1. In this embodiment, the determination of moving area/stationary area is performed on a block-by-block basis. The input is divided into blocks by a blocking circuit 202 and output. The frame memory 201 stores frame data of one frame before, and outputs block data at the same position of the previous frame data in accordance with the block data output of the blocking circuit 202.

A difference circuit 203 calculates a difference between the two block data, and a power calculation circuit 204 calculates the difference power for each block. This power P is a fixed threshold value t of the threshold value processing circuit 205.
It is compared with hl, and whether P>thl or P≦thl is output as 0/1. The map creation circuit 206 collects the 0/1 data for each block and outputs it as an O/1 map for the frame.

FIG. 3 shows a block diagram of another embodiment of a dynamic area detection circuit. In this embodiment, input data is compared with the background memory to determine whether a static area or a moving area is present, so that determination can also be made on a pixel-by-pixel basis.

FIG. 4 shows in detail one embodiment of the configuration of the background memory 301 shown in FIG. The frame removal control circuit 41 controls the input images of two adjacent frames to be sequentially transferred to the frame memories 42 and 43 in order to obtain a difference image between the two adjacent frames by the difference circuit 44, and controls the interval between the sets of the two frames. In order to obtain an appropriate value, frame replacement control is performed in response to a computation end signal from the update control circuit 49. Differential circuit 44
The inter-frame difference image of the two frames of input images in frame memory 42.4B determined by is stored in frame memory 46 after its absolute value is taken by absolute value circuit 45.

The 46 contents of this frame memory are input to a contour detection circuit 47, and a more general contour than that obtained by the map creation circuit 23 of FIG. 3 is detected.

The map creation circuit 303 creates a map showing updated positions as shown in FIG. 5 based on the output of the contour detection circuit 47. The update control circuit 49 refers to this map and controls whether to update the contents of the background frame memory 50 with the contents of the frame memory 42 or to save the contents of the background frame memory 50.

The delay circuit 51 includes a difference circuit 44 - absolute value circuit 45 - frame memory 46 - contour detection circuit 47 - map creation circuit 48
The output of the frame memory 43 is delayed by a time corresponding to the calculation time up to the update control circuit 49, and the update control circuit 49
Writing is performed in the background frame memory 50 in pixel units only when updating is permitted by a signal from the background frame memory 50. The operating principle of this background memory will be briefly explained below.

Figure 6 shows an example of the outline of a moving area (10) (in this example, a person) determined from the inter-frame difference image. ) and background parts that are newly visible compared to the previous input frame, so it is not possible to update the hidden background in one update.However, due to the movement of a person, etc. There is a possibility that the parts that were not updated will be updated from the second time onwards, and as shown in Figure 7, the background will be changed by updating the parts that are newly included outside the outline of the inter-frame difference image. It becomes possible to store only the background image in the background memory 24.

For this reason, as shown in Fig. 8, two frames of images that are temporally close to each other so that the area of the background area inside the outline is small are used as the two frames of images for which the inter-frame difference image is to be obtained. However, these two
The interval between the frame sets can be set arbitrarily depending on the desired update interval and calculation time. In FIG. 8, frames from t1+Δ1 to t2, from ti to t1+Δ1, and from t2 to t2+Δ2 are generally deleted by the frame deletion control circuit 41. However, Δ1 and Δ2 may be frame intervals. Here, an inter-frame difference image 1 is obtained from the difference between frame t1 and frame t1+Δ1, and an inter-frame difference image 2 is obtained from the difference between frame t2 and frame t2+Δ2. will be updated. In Fig. 7, the diagonally shaded area rising to the right is (tl,
1+Δ1), and the diagonal lined area downward to the right is (t2, t2+Δ1).
2) shows the updated parts between.

FIG. 9 shows an example of the configuration of the contour detection circuit 47 in FIG. 4. The blocking circuit 61 generates L as shown in FIG.
A rectangular block of XH (H is the length of one side of the frame) is taken out in the vertical and horizontal directions, and the histogram creation circuit 62 creates a histogram in the direction of the short side of the rectangle. The end point detection circuit 63 detects the histogram from both ends of the strip.
A search is performed while comparing with a certain threshold value, and the coordinates of each point that exceeds the threshold value for the first time are output as a contact point of the contour.

Next, FIG. 11 shows a method for separating a background and a moving area, which is performed by the edge detection circuit 302 in FIG. 3.

As shown in FIG. 3, the difference between the frame input and the contents of the background memory 301 is input to the edge detection circuit 302.

Detecting these contours often results in a region that includes the contour of the actual motion region.

Therefore, consider a band of a certain width inside this contour, and add an edge detection operator (for example, 5 obel operator) inside this band.
Then, the pixel with the larger result is searched within the band in the direction of the short side of the band and is used as a new outline.

The map creation circuit 303 outputs a map in which static/motion is converted to 0/1 based on the detection results obtained in FIG.

Next, Fig. 12 shows the signal distribution calculation circuit 103.10 of Fig. 1.
FIG. 4 shows a block diagram of an embodiment of No. 4. The input pixel data is input to the pixel value addition circuit 1201. The pixel power integration circuit 1202 calculates the cumulative addition value of the pixel value and the cumulative addition value of the square of the pixel value.
In step 03, the number of pixels in each area (moving area or static area) is counted. When the manual operation of the entire area is completed, the values counted by each circuit are sent to the average value calculation circuit 1204 and the variance calculation circuit 1205, where the average value and variance of the pixel values in the area are calculated and output. Pixel value addition circuit 12 inches,
The values of the pixel power integration circuit 1202 and the pixel number counter 1203 are reset at the timing of inputting pixel data within the area.

FIG. 13 is a block diagram showing another embodiment of a signal distribution calculation circuit. This circuit is a circuit that calculates the maximum and minimum values within the area. The input to each area is input in synchronization with the frame synchronization signal. (Input when frame synchronization) Comparator 1
301, the maximum value candidate held in the maximum value hold circuit 1303 and the input pixel value are input for each pixel, and when 1 manual search 1>1 maximum pixel 1, a hold EN signal is output and a new The pixel value input to the maximum value hold circuit 1303 is held. The held maximum value is output at the rising edge of the frame synchronization signal. Furthermore, the contents of the maximum value hold circuit 1303 are reset to φ at the next fall of the frame synchronization signal. The operations of the comparator 1302 and the minimum value hold circuit 1304 are also completely similar except that the reset value of the minimum value hold circuit is FF.

The explanation of the above two examples assumes that the static area/moving area can be completely separated pixel by pixel, but in the case of block unit separation, the static area and moving area are mixed in some blocks, and the maximum There may be cases where appropriate values such as values and minimum values cannot be determined. In such a case, when calculating the signal distribution in the moving region as shown in FIG. 16, it is sufficient to calculate the signal distribution in a region narrower than the obtained boundary, such as the shaded area in the diagram.

Next, the operation of the luminance conversion circuit 105.10[i in FIG. 1 will be explained. FIG. 14 is a diagram explaining the principle of operation. For example, in the case of backlighting, the brightness of the background is much higher than that of the person, and since there is less light on the person, the contrast is low, resulting in a signal distribution as shown in FIG. 14(a).

The principle is that by reducing the brightness of the background and increasing the brightness and contrast of the person as shown in FIG. 14(b), an image that is easy to see can be obtained. In this case, for example, the average brightness of the person and background in the original image,
Distribute MOP, MOB, VOP.

If it is a VOB, it is conceivable to convert the pixel value using the following formula. In such a case, it is conceivable that the brightness conversion circuit 105.10B of FIG. 1 is configured with a microcomputer or a multiplier to execute the ``1'' mathematical expression in software. ”-The mathematical expression may be classified into several types of conversion characteristics depending on the values of M and V. Also, visually, nonlinear conversion characteristics as shown in FIG. 15 may be better. In these cases, a configuration may be considered in which the outputs of the conversion characteristic ROMs 1 to ROMn are switched depending on the calculation result of the signal distribution, as shown in FIG. 17.

Next, another example of conversion characteristics will be explained using FIGS. 18 to 20.

FIG. 18 shows typical brightness that appears during backlighting. In order to widen the dynamic range of each region, it is sufficient to perform brightness conversion so that the frequency of appearance becomes flat in each region. Figure 19 shows the distribution in Figure 18 rewritten as a cumulative frequency. For such a distribution, the cumulative histogram is divided into equal intervals in the output axis direction as shown in Figure 20, and the corresponding input luminance interval is Brightness conversion is performed by input/output conversion that makes the inside correspond to the same output brightness.

As a result, the above-described conversion is performed. According to this method, even if the bright luminance of the background is erroneously counted in the histogram of the human area, the effect on the conversion is small. Furthermore, since the bias in the brightness distribution is also corrected, an easy-to-see image can be obtained regardless of the brightness distribution. And the histogram h (i) i
This can be easily realized using the transformation matrix Xfm, which is between -0 and 255. The above conversion expands the dynamic range and makes it easier to see, but since the dynamic resin, which is already small, is forcibly expanded, there is a possibility that the problem of quantization becoming rougher and granular noise increasing may occur. (Figure 21(a)). As a countermeasure for this, the precision of the input A/D conversion should be adjusted to the output quantization precision (for example, 8bit).
It is conceivable to create a histogram and perform luminance conversion with a finer resolution (for example, 12 bits) than t), and then round it back to the output quantization precision at the time of output (FIG. 21(b)).

An example of a circuit for performing this calculation is shown in FIG.

The input and detection map are input synchronously, and 110 of the map
Luminance histograms are created separately depending on the brightness. This is sent to the conversion characteristic determining circuit, and the conversion characteristic of each area is determined by (*). The input is stored in the memory, and after determining the conversion characteristics, it is sent to the conversion circuit again in synchronization with the map, where the conversion corresponding to 0/1 of the map is performed and output.

As another example, it is conceivable that the transformation characteristics are determined only for the human region, and the same transformation is applied to the background. In this case, the background may be converted to a high luminance value, making the screen difficult to see, so it is preferable to discontinue the conversion of the background above a certain value th, as shown in FIG. When converting moving images, if the conversion characteristics change from frame to frame, there is a possibility that undesirable flickering may occur. To avoid this, the conversion characteristics are determined only once every several frames, or the histogram shown in FIG. 22 is created over several frames. Flicker can be prevented by weighting and adding the histograms up to the previous frame and the histogram of the current frame. FIG. 24 shows a block diagram of a conversion circuit for flattening a histogram, which is different from that shown in FIG. 22. This circuit differs from the embodiment shown in FIG. 22 in that it performs conversion using conversion characteristics determined by the histogram of one frame before. First, at the time of inputting frame data, the change characteristics determined by the previous frame are stored in the conversion characteristic RAM. In this case, when a pixel value is input as an address, a converted value is obtained as an output, and luminance conversion is thereby performed and output. The pixel value is also input to the address of the histogram RAM at the same time, the frequency of the corresponding pixel value is read out, 1 is added by the adder 1, and it is written to the same address. By performing this operation for one frame period, a pixel histogram for one frame is created in the histogram RAM. At this time, the moving area detection result is input to a pixel number counter and a histogram RAM, and the number of image pixels within the moving area is counted. Also, the histogram is created only for the moving region. The pixel number counter and histogram RAM are cleared until the next frame data is input. After inputting the frame data, the conversion characteristics are determined while reading the histogram from the histogram RAM.

The histogram read from the histogram RAM is converted into a cumulative histogram by a combination of an adder 2 and a flip-flop. The division circuit divides this cumulative histogram by the number of pixels in the moving area, and the multiplication circuit multiplies it by 255 and writes it into the conversion characteristic RAM, thereby preparing the conversion characteristic for the next frame. At this time, it is sufficient to input addresses that change sequentially from 0 to 255 to the addresses of the conversion characteristic RAM and the histogram RAM.

(I9) In addition, a determination circuit may be provided as shown in Figure 25 to determine from only the signal distribution of the signals within the frame that there is a light bulb or fluorescent light within the screen and the face is dark. can. In this case, a certain threshold value (for example, TL
-254) Check the number of pixel values that are greater than the second threshold TN, e.g. TN=80);
It is determined that this is the brightness caused by a light bulb or fluorescent lamp, and as a result, the brightness conversion circuit operates. Otherwise it won't work.

Note that malfunctions can easily be specified by the user as desired.

As described above, brightness conversion can be automatically performed by digital signal processing of an input image. Furthermore, since images can be input using a camera with a simple structure that does not have functions such as an aperture, it is also possible to reduce the size and cost of the moving image transmission device. Furthermore, nonlinear processing can be performed on the input image, and a brightness correction method that is easier to see can be selected.

In addition, instead of correcting the brightness of the input image,
As shown in FIG. 26, it is also possible to adjust the aperture of the camera using the signal distribution of the moving area detected by the moving area detection circuit and the signal distribution calculation circuit. and,
If the signal distribution in the moving area on the screen is narrow, it may be because the face or the like is darkened by a light bulb or fluorescent light, so the camera aperture can be controlled to open further.

Further, as shown in FIG. 27, the brightness may be adjusted at an analog level by adjusting the gain of the video amplifier after inputting the image. When a face or the like is dark and processed by an analog signal circuit, the signal level is controlled to increase. Note that since there is a limit to the signal level, a limiter may be used.

Furthermore, although the signal amplification may be symmetrical in positive and negative directions, it is also possible to have an asymmetrical amplification characteristic, for example, increasing the expansion only in the positive direction.

Note that the detection method may be performed using methods other than inter-frame differences. This is, for example, a distance measurement method using stereo images.

In addition to processing the image input by separating it into a moving region and a still region as shown in FIG. 1, it is also possible to adopt a simplified configuration as shown in FIG. 28.

FIG. 28 assumes that a person is captured in the center of the screen, and as shown in FIG. 29, the signal distribution is calculated only in a fixed luminance conversion area provided at the center of the screen. This method uses the results to transform only the area.

With this method, it is not necessary to perform dynamic area detection. The assumed motion area may be fixed as shown by the dotted line in FIG. 29, or only the size of the area may be fixed and only the starting point may be based on the detection result. According to this method, by fixing the size in the case of histogram conversion, the 24th
A counter is no longer required for the number of pixels shown in the figure. Further, by setting the size to a power of 2, the division at the time of creating the histogram can be done by bit shifting, so the circuit scale can be reduced.

As an example of extracting a fixed brightness conversion area, FIG. 30 shows a case where a spot photometry detector is provided in addition to the camera. This means using an optical detector with a narrower field of view than the input camera. This spot photometry detector performs spot photometry only on the center of the screen as shown in FIG. 31, and the gain of the camera's video amplifier is controlled based on this result. Therefore, extraction of a brightness conversion area and brightness correction can be performed with a simple configuration. Note that if the direction of the spot photometry detector is configured to be freely changeable, the user can move it to respond even if the other party is not in the center of the screen.

As yet another embodiment, as shown in FIG.
A window generation circuit generates a window signal corresponding to the width (measured value) of the face on the screen in the drive circuit. It is also possible to accumulate only the CCD signals within the window, measure their levels, and determine the gain of the video amplifier based on the measurement results. Note that the width of the window signal shown in FIG. 33 can be changed by the user according to the width of the face on the screen.

Next, a case where a color signal is added to the input image signal will be described.

FIG. 34 is a block diagram of a moving image transmission apparatus in which an image is input as an input image signal. The input image signal is first subjected to region detection as a rectangular block including a face in a face detection circuit 1801 and sent to a skin detection circuit 1802 . Excluding particularly dark areas such as hair and mouth from the center of the face, the skin area is detected starting from the center. The signal distribution information of this skin region includes a maximum value, a minimum value, an average value, etc., but for example, when an average value is used, a conversion is obtained so that this average value becomes a standard skin color. The X mark in Figure 35 is this detected average value, and the * mark is the standard skin color.To convert from × to *, for example, add the vector from All you have to do is limit it.

As a result, by changing the color tone of the screen so that the human skin area has the correct color, the color tone can be approximated to the correct color even when using the device in a dimly lit area.

Further, in this embodiment, a determination circuit 1805 determines the skin area detection result to determine whether or not the detection was appropriate. This determination can be made, for example, based on the size, signal content, shape, etc. of the detection area. If the detection is determined to be correct, the same operation as in the above embodiment is performed, and if not, the color information conversion is stopped. Furthermore, this switching can also be performed using the manual switch 1806. Furthermore, when it is desired to fix the parameters that define the conversion method of this color information conversion, it is sufficient to similarly input preset instruction information using a manual switch or the like.

In addition, the color tone will differ depending on whether the lighting is a fluorescent lamp or incandescent lamp, natural light from the sun or infrared light, but by correcting the edges of the spectrum of the light source that is illuminating, a more natural color tone can be displayed. be able to. Light sources with special spectra, such as fluorescent lamps, can be automatically detected and corrected. Alternatively, the user may specify the light source and change the correction method based on the user's settings.

As an example of a correction method, the color signal is set to 25 around 0.
) (26) There is a method in which only the enlargement and reduction conversion is performed and the average value is not added.

Although the case of color information conversion has been described, it is also possible to perform the above-mentioned luminance change.

Also, when correcting brightness and color tone by signal processing,
It is also possible to set the camera so that the image is brighter than usual and the color tone is saturated when the image is input by the camera. All of the above explanations were examples that were performed at the time of input on the sending side, but
The same thing can also be done as output signal processing on the receiving side. FIG. 36 is a block diagram showing an example of the configuration of the receiving side.

In this case, if the encoding method uses conditional pixel replenishment, the transmitted block is a block in the moving area, so
A similar conversion can be performed by determining the signal distribution from these. Further, as another embodiment, the same method can be used in which the area is fixed (same as the embodiment shown in FIG. 29).

[Effects of the Invention] As explained above, according to the moving image transmission device of the present invention, image correction is automatically performed to make it easier to see even images shot under poor lighting conditions. Image quality can be improved because correction is performed with points placed on the human area.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of a moving image transmission device according to the present invention, FIGS. 2 and 3 are block diagrams showing an embodiment of the moving area detection circuit of FIG. 1, and FIG. FIG. 3 is a block diagram showing an example of the configuration of the background memory; FIG. 5 is a block diagram showing an example of the map in the background memory shown in FIG. 4; FIGS.
The figure shows the principle of updating the background memory in Fig. 4, Fig. 9 shows an example of the configuration of the contour detection circuit in Fig. 4, Fig. 10 shows the principle of contour detection, and Fig. 11 are diagrams showing an example of how to obtain the contour in the separation circuit of Figure 1, Figure 12,
Fig. 13 is a diagram showing an example of the configuration of the signal distribution calculation circuit in Fig. 1, Fig. 14 is a diagram showing the principle of luminance conversion in the luminance conversion circuit in Fig. 1, and Fig. 15 is an example of luminance conversion characteristics. FIG. 16 is a diagram showing the ideas for calculating signal distribution, FIG. 17 is a diagram showing an example of the configuration of a luminance conversion circuit, and FIG.
Figures 2 to 20 are diagrams showing examples of other luminance conversion characteristics.
Figure 1 is a diagram used to explain quantization accuracy, Figures 22 and 23 are block diagrams of a circuit for determining conversion characteristics by flattening the histogram, and Figure 24 shows the conversion characteristics determined by the histogram of one frame before. A block diagram of the conversion characteristic determination circuit to be used, FIG. 25 is a diagram in which a determination circuit is added to the signal distribution calculation circuit, FIGS. 26 and 27 are diagrams showing examples other than correcting the brightness of the input image, Figure 28-3
FIG. 3 is an explanatory diagram when image input is simplified, and is a diagram showing an example of the configuration of the receiving side of the moving image transmission device.

Claims (4)

[Claims] (1) a transmission means for transmitting an input moving image; a moving area detecting means for detecting a moving area from the input moving image; a brightness distribution calculating means for calculating a brightness distribution of the moving area detected by the detecting means; A moving image transmission device comprising: means for adjusting the brightness of the input moving image according to a calculation result by the brightness distribution calculating means. (2) a transmission means for transmitting an input moving image; a moving area detecting means for detecting a moving area from the input moving image; a brightness distribution calculation means for calculating a brightness distribution of the moving area detected by the detecting means; A moving image transmission device comprising: brightness converting means for converting the brightness of the input moving image according to a calculation result by the brightness distribution calculating means. (3) A transmission means for transmitting an input moving image, a moving area detecting means for detecting a moving area from the input moving image, and a brightness distribution of the moving area detected by the detecting means and an area other than the moving area. comprising a brightness distribution calculation means, and a brightness conversion means for performing brightness conversion of the input moving image according to a calculation result by the brightness distribution calculation means, and the brightness conversion means performs brightness conversion of the moving area and other than the moving area. A moving image transmission device characterized in that transmission is performed separately. (4) a transmission means for transmitting an input image, a detection means for detecting a skin area of a person from the input image, and a brightness distribution calculation means for calculating at least the brightness distribution of the skin area by this detection means; A moving image transmission device comprising: color information conversion means for converting color information of the skin area using the brightness distribution calculated by the brightness distribution calculation means.