KR100215206B1 - Video conference system - Google Patents

Abstract

The present invention relates to a video conferencing system for recognizing a conversant using an infrared image and then tracking it.

The present invention relates to a method and a system for recognizing a model, comprising: a model recognition step of extracting and recognizing contours of respective model faces by inputting respective model image signals corresponding to infrared images of one or more talkers during a video conference; A model selection step of selecting one of the recognized one or more model image signals; A talker recognizing step of extracting and recognizing a contour line of a talker's face from an infrared image signal of a talker input through a camera; A matching step of matching a contour of a model face of the selected model video signal with a contour of a talker face of the recognized talker video signal; And an interpreter tracking step of tracking the matched talker and moving the camera, wherein each model image signal corresponding to the infrared image of one or more talkers during a video conference and the infrared image signal of the talker are input, Recognition means for extracting and recognizing contours of the model and the face of the talker; A CPU for controlling the operation of the recognizing means and for controlling the extracted model video signal to match the outline of the face of the talker and to track the registered talker; And a motor driver for moving the direction of the camera under the control of the CPU.

Accordingly, the present invention has a function of recognizing a talker using an infrared image and tracking the talker, so that even if a talker moves during a video conference, it can be tracked by a camera.

Description

VIDEO CONFERENCE SYSTEM

The present invention relates to a video conference system, and more particularly, to a video conference system having a recognizer function.

It is an object of the present invention to provide a video conferencing system for allowing a conversation person to freely move by tracking a conversation person even if the conversation person moves without being positioned in front of the camera.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video conferencing system, and more particularly, to a video conferencing system for recognizing a conversant using an infrared image and tracking the same.

In general, a video conferencing system allows a conference attendee to stay on the screen, even if they are at a distance.

Therefore, the dialogue person must be located in front of the camera, so that he can show his / her appearance to the remote talker who is remotely through the screen.

That is, the conventional video conferencing system has a problem that the conversation person can not move because the conversation person must be located in front of the camera at the time of video conferencing.

An object of the present invention is to provide a video conferencing system for recognizing a conversation person using an infrared image and tracking the conversation person to freely move the conversation person.

1 is a flowchart showing a model recognition step according to the present invention.

Figure 2 is a flow chart illustrating the recognizer, registration and tracking steps of the present invention,

FIG. 3A is a view for explaining the boundary extracting steps of FIGS. 1 and 2;

FIG. 3B is a view for explaining the thinning step of FIGS. 1 and 2;

Fig. 3C is a view for explaining the linear approximation steps of Figs. 1 and 2

4 is a detailed flowchart showing the tracking step of FIG. 2;

5 is a block diagram of the video conferencing system according to the present invention

DESCRIPTION OF THE REFERENCE NUMERALS

400: infrared camera 500: recognition unit

510: A / D converter 520: memory

530: boundary extracting unit 540: boundary line calculating unit

550: Fine line section 560: Line approximation section

570: Linear feature extraction unit 600: CPU

610: ROM 700: motor driver

800: Motor

In order to achieve the above object, a video conference system according to the present invention includes a model recognition step of extracting and recognizing contours of respective model faces by inputting respective model video signals corresponding to infrared images of one or more talkers during a video conference, ; A model selection step of selecting one of the recognized one or more model image signals; A talker recognizing step of extracting and recognizing a contour line of a talker's face from an infrared image signal of a talker input through a camera; A matching step of matching a contour of a model face of the selected model video signal with a contour of a talker face of the recognized talker video signal; And a communicator tracking step of tracking the matched talker to move the camera, wherein each model image signal corresponding to one or more infrared images of one or more talkers during a video conference and an infrared image signal of the talker are input And recognizing means for extracting and recognizing the contour of the face of the talker; A CPU for controlling the operation of the recognizing means and for controlling the extracted model video signal to match the outline of the face of the talker and to track the registered talker; And a motor driver for moving the camera according to the control of the CPU.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 and 2, the video conference system according to the present invention includes a model recognition step 100, a model selection step 200, a talker recognition step 201 to 207, a matching step 208 and 209, , The talker tracking step 210, and the model change step 211, 212.

As shown in FIG. 1, the model recognizing steps 100 to 107 extract the outline of each model face by inputting respective model image signals corresponding to infrared images of one or more talkers during a video conference, A model image signal storing step (100) of A / D converting and storing a model image signal corresponding to an infrared image of a talker to be recognized at the time of a video conference, a boundary extracting step of calculating a boundary pixel and a boundary direction from the stored model image signal A step of thinning the extracted boundary pixels by using the extracted boundary direction, a step of extracting a linear feature of a straight line by tracing the thinned boundary pixel, A linear linear feature extraction step (104, 105, 106) for storing a linear model feature value, and another model image signal corresponding to an image of another talker to be recognized And if so, proceeds to the model image signal storage step 100. [0041] FIG.

Here, the boundary pixels are determined by comparing vertical and horizontal differential values of the stored model video signal with a threshold value for setting a boundary, and the boundary direction is determined by comparing arctangent of the vertical and horizontal differential values of the boundary pixel Value (aran (h / v)).

The model selection step 200 is a step of selecting one of the recognized one or more model video signals.

As shown in Fig. 2, the talker recognizing steps 201 to 207 are steps of extracting and recognizing the outline of the talker's face from the infrared image signal of the talker input through the camera, A boundary image extracting step (202, 203) for calculating a boundary pixel and a boundary line direction from the stored intermittent image signal, an extracting step (202, 203) for extracting an infrared image signal of the talker from the stored intermittent image signal, And a linear linear feature extraction step (205) of extracting and storing a straight line linear feature along the thinned boundary pixel and storing the linear feature, 206, and 207, respectively.

Here, the boundary pixels are determined by comparing vertical and horizontal differential values of the stored communicator image signals with thresholds for discriminating boundary lines, and the boundary line direction is determined by comparing arc tangent values of the vertical and horizontal differential values of the boundary pixels (aran (h / v)).

The matching steps 208 and 209 match the outline of the model face of the selected model video signal and the outline of the talker face of the recognized talker image signal.

The talker tracking step 210 is a step of tracking the matched talker to move the camera. As shown in FIG. 6, a predetermined size window including the talker's face is set at an initial position where the matched talker is located A window setting step 300, a binarization step 301 for binarizing pixel values in the set window, a threshold setting step 303 for setting a threshold value for tracking the center point of the matched talker, a threshold value setting step 303 for setting the binarized pixel value to a set threshold value A camera moving and tracking step (step 304) of moving the camera and performing tracking if the binary value of the binarized value search result is greater than the threshold value, and a camera moving and tracking step If it is impossible to trace the talker during the execution of the talker step 304, the process proceeds to the talker recognition step 201 to 207, It is performed by the formula and not re-trace trace processing stage (305, 306).

The model changing steps 201 to 209 are steps 211 and 212 of moving the camera and proceeding to the camera recognition and matching steps 201 to 209 when selecting one of the recognized one or more model video signals, to be.

Detailed operation of the video conferencing system according to the present invention will be described with reference to FIGS. 1 to 5. FIG.

First, an infrared model video signal corresponding to all talkers participating in a video conference should be taken through the camera to extract the face contours of the model.

That is, by performing the model recognizing steps 100 to 107, the contour lines of each model face are extracted and recognized by inputting respective model image signals corresponding to the infrared images of one or more talkers as a reference, Will be described with reference to the drawings.

First, a model video signal corresponding to an infrared image of a talker to be recognized at the time of a video conference is taken through an infrared camera and A / D converted and stored (100).

A signal received from an infrared camera, for example, an NTSC signal, is converted into a digital image I (x, y) having a value within a specified range on a two-dimensional matrix after A / D conversion. This image is stored in memory and used as input image.

After the model image signal storage step 100 is performed, boundary extraction steps 101 and 102 are performed to calculate boundary pixels and boundary lines from the stored model image signals.

(H, v) in the horizontal and vertical directions by differentiating the stored model video signal in the vertical and horizontal directions, and then calculates the difference (h ² + v ² ) ^1/2 of the differential value The boundary pixels are found by comparing threshold values.

A digital image I (x, y) signal having a value within a specified range on a two-dimensional matrix is differentiated in the horizontal and vertical directions as shown in the following equations (1) and (2).

I (x + 1, y-1) + I (x-1, y) + I -I (x + 1, y-1) = h

----- Equation (1)

I (x, y + 1) -I (x, y + 1) -I (x + 1, y + 1) = v

----- Equation (2)

As shown in FIG. 3, the horizontal differential includes the brightness values A4, A5, and A6 of the left pixel and the brightness values A1, A2, and A3 of the right pixel with the reference pixel x, A7 and A1 on the upper side and the brightness values A6, A8 and A3 on the lower side are centered on the reference pixel x and y, It is done by retrieving the car.

As shown in the above equation (1), the horizontal differential value is calculated to calculate the horizontal differential value h and the vertical differential value is calculated as shown in the above equation (2) to calculate the vertical differential value (v) (H ^2, v ² ) ^1/2 ) of the differential values (h, v). That is, the magnitude ((h ² + v ² ) ^1/2 ) of the differential value is calculated by squaring and adding the vertical and horizontal differential values (h, v), and then taking the square root again.

When the calculated differential value ((h ² + v ² ) ^1/2 ) is compared with a threshold value for determining a boundary and is larger than the threshold value for threshold value determination, the reference pixel (x, y) becomes a boundary pixel.

The size of the differential value and the comparison process are performed for each pixel of the model video signal to extract all boundary pixels in the model video signal of one frame.

After extracting the boundary pixels and extracting the boundary lines as described above, the boundary direction is calculated (102), and the boundary direction is calculated as the arctangent value (aran (h / v)) of the differential values in the vertical and horizontal directions of the boundary pixel .

That is, if a look-up table is prepared in the internal ROM, an arctangent result for the value (h / v) can be obtained.

After extracting the boundary pixels and the boundary direction, the extracted boundary pixels are thinned using the extracted boundary direction (103).

That is, since the boundary pixels extracted by differentiation are not thicker than one pixel but have a thickness of about 2 to 3 pixels due to the characteristic of the infrared image, the boundary pixels are thinned to extract more accurate boundary pixels.

This will be described with reference to FIG.

The boundary pixels are checked with respect to the boundary line and the boundary pixels are excluded except for the one pixel having the largest boundary pixel value.

That is, by checking the pixels of the boundary pixel (e) direction boundary line (a) and the direction (b) which is 90 degrees as shown in the 4 magnitude of the differential value ^{(h 2 + v 2) 1/2} ) is most Leaving only the large boundary pixels and excluding the remaining boundary pixels.

After the thinning step 103 is performed, a linear linear feature extraction step 104, 105, or 106 is performed to linearly approximate the thinned boundary pixel and extract and store a linear feature of a straight line.

5, if the distance between the pixel e and the straight line l1 is equal to or larger than a certain distance by starting the straight line 11 from the current pixel e1 along the boundary pixel e , One straight line (11) ends and another straight line (12) starts to be linearly approximated.

After extracting the approximated straight lines (l1, l2) along the boundary pixels, the linear features of these straight lines are extracted (105). That is, when the extracted approximated straight lines are n, (Straight line 1, start point, end point, length, slope), (straight line 2, start point, end point, length slope), ..., (straight line n, slope, end point, length slope).

The number of straight lines that can be extracted from the contour of the face of the input image varies depending on the image, but the number of straight lines (n) that can be extracted per person's face in a video conference is about 10 to 20.

Thus, the linear feature extracted from the model video signal is used for matching with the linear feature extracted from the talker image signal stored in the memory.

After extracting the linear feature of the model image signal of one model, it is determined whether there is another model image signal corresponding to the image of the other talker to be recognized. At this time, if there is another speaker to be present, if there is another model video signal, the process goes to the model video signal storing step 100 to extract a linear feature for another model video signal. Also, if a linear feature is extracted for all the model video signals and there is no other model video signal, a model selection step 200 for terminating the model recognition step and selecting one of the recognized one or more model video signals is performed do.

That is, a model to which a video signal is to be transmitted is selected through a camera (200).

For example, one of the many talkers attending a video conference mainly selects one talker to talk to.

After selecting the model in this way, it is necessary to find and track the same talker as the selected model among attendees who attend the video conference.

Accordingly, as shown in FIG. 2, a talker recognizing step (201 to 207) for extracting and recognizing the outline of the talker's face from the infrared image signal of the talker input through the camera is performed.

The talker recognition step is performed in the same manner as the above model recognition step.

That is, as shown in FIG. 2, a speaker signal processing step 201 for A / D-converting an infrared image signal of a talker input through a camera and storing the signal is first performed. Then, A boundary extracting step (202, 203) for calculating the boundary line direction is performed.

At this time, the extracted boundary pixels and the boundary direction are performed in the same manner as the boundary extraction step performed in the above model recognition step.

That is, as shown in Fig. 3, the digital video signal I (x, y) of the talker A / D-converted and having a value within a specified range on the two- (H, v) in the horizontal and vertical directions are differentiated in the vertical and horizontal directions, and then the magnitude ((h ² + v ² ) ^1/2 ) of the differential value is compared with the threshold value for setting the boundary, (202).

The boundary direction is calculated from the arc tangent value aran (h / v) of the vertical and horizontal differential values of the extracted boundary pixels (203).

In addition, as in the model recognition step, a thinning step 204 for thinning the extracted boundary pixels using the calculated boundary direction is performed as shown in FIG. 4, Linear linear feature extraction steps (205, 206, 207) for extracting and storing straight line approximated linear features are performed as shown in FIG.

The linear feature of the talker extracted and extracted from the talker image signal is extracted from the model image signal and then compared with the linear feature of the selected model to determine whether the match is true or false.

That is, the linear feature of the model face image of the selected model image signal is compared with the linear feature of the contour face of the recognized face image of the recognized speaker image signal to determine whether or not the matching is performed. ) Method.

That is, a straight line of all models that can be matched with respect to the straight line of the input image is assigned, and then a specific value m is calculated as shown in the following equation (3), and it is considered that the matching is performed when the value m is the smallest.

n

m = 1 / (? # (input image straight line length i - model straight line length #

i

+ # (Input image straight line slope i - Model straight line slope # () - (3)

If the matching is not performed, the camera is moved to select another talker, and the talker image signal of the other talker is input repeatedly from the talker image signal storing step 201 to be selected Find a talker with the same linear characteristics as the model.

After finding the communicators that match with the model through the above process, the tracking is performed through the center point tracking method.

That is, if the matching is performed, the initial position can be estimated for the talker, so that the initial information is used to move and track the camera 210, which will be described in detail with reference to FIG.

First, a window setting step 301 is performed to set a window of a predetermined size including the face of the talker at an initial position where the matched dialogue is present.

Generally, the center point tracking is performed on the entire image. However, in the present invention, since the initial position can be known through the recognition function, the tracking function is limited to the vicinity of the initial position. That is, a window that is a window of a size including a face of a person can be set to perform the tracking function only in this window.

Here, the size of the window may be a size that includes the size of the face in the image but does not include the face of another person excluded from the recognizer.

After the window setting step 300 is performed, a binarization step (step 301) of binarizing pixel values in the set window and a threshold value setting step of setting a threshold value Thres for tracking the center point of the matched talker 302).

Since the face part can be easily extracted by setting the threshold value in the window, the binarization value searching step is performed as shown in the following equation (4) by setting the threshold value.

I (x, y) = 1 if I (x, y) Thres --- (4)

In the above equation (4), if the binarization value is retrieved and the value of the image in the image is '1', tracking of the talker becomes possible. In other words, when the motor is driven so that the camera moves to the portion where the value of '1' is indicated by the initial recognition and the part indicated by '1' moves together with the conversation person, the movement information is transmitted to the motor again to move the camera (304).

Therefore, since the window is moved with the camera, it becomes possible to keep track of it.

On the other hand, when two or more faces are photographed in the sudden movement of the target face or the window during the movement of the camera for tracking, the recognition target face must be recognized again through recognition function again by activating the recognition function.

That is, if it is impossible to trace the talker during the camera movement and tracking step 304, the process proceeds to the talker recognition steps 201 to 207 to perform the traceability processing steps 305 and 306 for re-recognizing and re-tracing the talker .

Next, the video conference system is composed of the recognition unit 500, the CPU 600, and the motor driver 700 as shown in Fig.

The recognition unit 500 extracts and recognizes contours of faces of each model and talker by inputting respective model image signals corresponding to infrared images of one or more talkers during a video conference and infrared image signals of the talker, An A / D (Analog / Digital) converter 510 for A / D-converting each model image signal and a talker image signal corresponding to infrared images of one or more talkers under a control of the CPU 600, A memory 520 for storing a video signal outputted from the A / D converter 510 under the control of the CPU 600, a boundary extracting unit 520 for extracting a boundary pixel from a video signal outputted from the memory 520, A boundary line direction calculating unit 540 for calculating a boundary line direction using the boundary pixels extracted by the boundary extracting unit 530, a boundary line calculating unit 540 for calculating a boundary line direction calculated by the boundary line calculating unit 540, A linear approximation unit 560 that linearly approximates the boundary pixels along the thinned boundary pixels in the thinning unit 550, and a linear approximation unit 560 that linearly approximates the boundary pixels along the thinned boundary pixels, And a linear linear feature extraction unit 570 for extracting a linear feature of a straight line output from the linear approximation unit 560 and storing the extracted linear feature in the memory 520.

The CPU 600 includes a ROM 610 and controls the operation of the recognition unit 500 to control the matching of the extracted model video signal and the outline of the face of the talker and tracking of the registered talker.

The motor driver 700 drives the motor 800 to move the direction of the camera under the control of the CPU 600.

The operation of the hardware of the video conferencing system according to the present invention constructed as above will be described.

First, a signal from the infrared camera 400 is stored in a memory to generate an input image. A signal input from an infrared camera, for example, an NTSC signal, is converted into a digital image I (x, y) having a value within a specified range on a two-dimensional matrix through an A / D converter 510. This image is stored in a memory (520) and used as an input image.

After the digital image is stored in the memory 520, the function of storing the signal in the memory 520 from the A / D converter 510 until the operation of the linear feature extraction unit 570 is completed by the CPU 600 Off so that the input image does not change.

Of course, when the linear feature extraction operation of the linear feature extraction unit 570 is completed, the A / D converter 510 is turned on so as to receive a new input image again.

The boundary pixels are extracted by the derivative in the boundary pixel extracting unit 530. That is, the image signals I (x, y) outputted from the memory 520 are extracted by the above equations (1) and As shown in FIG. 3, after the differential values h and v are obtained by performing the horizontal and vertical differential calculations, the magnitude ((h ² + v ² ) ^1/2 ) of the differential value is calculated again and the threshold value And finds a boundary pixel.

In addition, the boundary line direction calculating unit 540 calculates the boundary line direction by arctangent values aran (h / v) of the differential values in the vertical and horizontal directions of the extracted boundary pixels.

As shown in FIG. 4, the fine line section 550 checks the boundary line direction calculated by the boundary line direction calculating section 540 and the pixels that are 90 degrees out of the boundary pixels extracted by the boundary extracting section 530, Excluding the largest pixel, the remaining pixels are excluded to exclude the boundary pixels.

As shown in FIG. 5, in the line approximation unit 560, the thinned pixels in the thin line unit 550 are linearly approximated while following the boundary pixels, and the linear feature of the straight line is extracted from the linear feature extraction unit 570 And extracted.

Linear features of a straight line include [(straight line 1, start point, end point, length, slope), (straight line 2, start point, end point, length slope), ... , (Straight line n, slope, end point, length slope)].

The linear characteristic of the straight line for the extracted model is stored again in the memory 520. [

A model in which the face of one or more talkers is photographed with an infrared camera in advance is extracted as a linear feature and constructed as a database.

On the other hand, the linear feature should be extracted through the above process for the infrared image signal of the talker inputted through the current camera. That is, the talker image signal is stored in the memory 520 via the A / D converter 510 and then transmitted to the boundary extracting unit 530, the boundary line direction calculating unit 540, the thinning unit 550, the linear approximation unit 560, , And a linear feature extracting unit 570 extracts a linear feature for a contour around the face.

The extracted linear features are stored in the memory 520. At this time, the same talker as one selected from a plurality of models is extracted from the matching through the CPU 600, and the extracted talker is sent to the CPU 600).

As described above, the video conferencing system according to the present invention has a function of recognizing a conversant using an infrared image and tracking the conversant, so that the conversation can be tracked by a camera even if the conversant moves during a video conference.

Claims (12)

A model recognizing step (100 to 107) for extracting and recognizing contours of respective model faces by inputting respective model image signals corresponding to infrared images of one or more talkers during a video conference; A model selection step (200) for selecting one of the recognized one or more model image signals; A talker recognizing step (201 to 207) for extracting and recognizing the contour line of the talker's face from the infrared image signal of the talker inputted through the camera; A matching step (208, 209) for matching the contour of the model face of the selected model video signal with the contour of the talker face of the recognized talker image signal; And (210) tracking the matched talker to move the camera. ≪ Desc / Clms Page number 19 > The method of claim 1, further comprising: a model changing step (211, 212) for moving the camera and proceeding to the camera recognizing and matching step (201 - 209) when selecting another one of the recognized one or more model video signals Wherein the video conference system further comprises: 2. The method according to claim 1, wherein the model recognition step (100 to 107) A model image signal storing step (100) for A / D converting and storing a model image signal corresponding to an infrared image of a talker to be recognized at the time of a video conference; A boundary extracting step (101, 102) for calculating a boundary pixel and a boundary line direction from the stored model video signal; A thinning step (103) of thinning the extracted boundary pixels using the extracted boundary direction; A straight line feature extraction step (104, 105, 106) for extracting and storing a linear feature of a straight line while following the thinned boundary pixel; And And determining whether there is another model video signal corresponding to the video of the other talker to be recognized, if it is determined that there is another model video signal corresponding to the video of the other talker to be recognized. 4. The apparatus of claim 3, Wherein the threshold value is determined by comparing a differential value in the vertical and horizontal directions of the stored model video signal with a threshold value for setting a boundary. 5. The method of claim 4, And arctanent values aran (h / v) of the differential values in the vertical and horizontal directions of the boundary pixels. 2. The method according to claim 1, wherein the talker recognition step (201 to 207) A savant image signal storing step (201) for A / D-converting the infrared image signal of a talker inputted through a camera and storing the result; A boundary extracting step (202, 203) for calculating a boundary pixel and a boundary line direction from the stored talker image signal; A thinning step (204) of thinning the extracted boundary pixels using the extracted boundary direction; And And a linear feature extraction step (205, 206, 207) for extracting and storing a linear feature of a straight line while keeping track of the thinned boundary pixels. 7. The apparatus of claim 6, Wherein the predetermined threshold value is determined by comparing a differential value in the vertical and horizontal directions of the stored communicator video signal with a threshold value for setting a boundary. 8. The method of claim 7, And arctanent values aran (h / v) of the differential values in the vertical and horizontal directions of the boundary pixels. 2. The method of claim 1, wherein the talker tracking step (210) A window setting step (300) of setting a window of a predetermined size including a face of the talker at an initial position where the matched dialogue is located; A binarization step (301) of binarizing pixel values in the set window; A threshold setting step (303) of setting a threshold for tracking the center point of the matched talker; A binarization value search step (303) for searching whether the binarized pixel value is larger than a threshold value set; And And a camera moving and tracking step (304) for moving the camera and performing tracking if the binarized value is greater than the threshold value. 10. The method of claim 9, wherein the tracking step (210) If it is impossible to trace the talker during the camera movement and tracking step (304), the process further proceeds to the talker recognition step (201 to 207), and the non-traceable processing step (305, 306) Wherein the video conference system comprises: Recognition means (500) for extracting and recognizing contours of faces of respective models and talkers by inputting respective model image signals corresponding to infrared images of one or more talkers during a video conference and infrared image signals of talkers; A CPU 600 for controlling the operation of the recognizing means 500 and controlling the extracted model video signal to match the outline of the face of the talker and to track the aligned talker; And And a motor driver (700) for moving the direction of the camera under the control of the CPU (600). 12. The apparatus of claim 11, wherein the recognition means (500) An A / D (Analog / Digital) converter 510 for A / D-converting each model image signal and a talker image signal corresponding to infrared images of one or more talkers under a control of the CPU 600, ; A memory 520 for storing a video signal output from the A / D converter 510 under the control of the CPU 600; A boundary extracting unit 530 for extracting boundary pixels from the video signal output from the memory 520; A boundary line direction calculating unit 540 for calculating a boundary line direction using the boundary pixels extracted by the boundary extracting unit 530; A nebulizer 550 for nebulizing the boundary pixels extracted by the nebulae extractor 530 using the boundary line direction calculated by the boundary line direction calculator 540; A linear approximation unit 560 that linearly approximates the boundary pixels along the thinned boundary pixels in the thinning unit 550; And And a straight line feature extraction unit (570) for extracting a linear feature of a straight line output from the linear approximation unit (560) and storing the linear feature in the memory (520).