JP3399995B2 - Video signal tracking method and video camera system - Google Patents

Description

Detailed Description of the Invention

[0001]

[Table of Contents] The present invention will be described in the following order. Industrial applications Conventional technology Problems to be Solved by the Invention Means for Solving the Problems (FIGS. 1 to 14) Action (Figs. 1-14) Example The invention's effect

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal tracking method and a video camera system, and more particularly to a human face which can be displayed as an image having substantially the same size.

[0003]

2. Description of the Related Art Conventionally, there is a single-lens reflex camera as a camera having an auto-zoom function, and a CCD (charge coupled d
(evice) A system has been proposed in which a line sensor is used to accurately obtain the distance between a subject and a camera and a zoom motor is moved. In this conventional auto-zoom camera system for a single-lens reflex camera, the subject is limited to humans, and the user touches the face with the operation buttons.
The bust shot and whole body shot are selected in advance, and the zoom motor is driven so as to match the zoom lens position while referring to the zoom lens position with respect to the distance to the subject that is programmed in advance for each shot. Has been done.

In such a single-lens reflex camera, since a CCD line sensor is additionally provided for autofocusing, it is possible to judge the front pin and the rear pin and to measure an accurate distance, which is relatively simple. The auto zoom system can be configured.

[0005]

However, since the method of the auto zoom system of the single lens reflex camera cannot measure the accurate distance in the video camera having the hill climbing control type auto focus system, it cannot be used as it is. Is not available. The present invention has been made in view of the above points, and when a person as a subject approaches or moves away from the camera, it is possible to always obtain an image of a face of almost the same size. An attempt is made to propose a method and a video camera system.

[0006]

In order to solve such a problem, in the present invention, an output for displaying an image including a subject image portion on a display screen PIC based on an image pickup signal S1 obtained from an image pickup means 1 performing a zooming operation. Video signal S3
In the video signal tracking method and the video camera system for forming the image, the video to be displayed on the display screen PIC is divided into a plurality of small areas AR (SP11), and based on the color characteristics of the image pickup signal S1 corresponding to the small areas AR. Then, the subject image candidate small areas ARA1 and ARB1 that are candidates for the subject image portion are extracted, and when a blank small area occurs between the collection of the subject image candidate small areas ARA1 and ARB1, the subject image among the blank small areas is extracted. The blank small area adjacent to the candidate small areas ARA1 and ARB1 is defined as the subject image candidate small areas ARA1 and ARB.
1 and the detection areas ARA2 and ARB of the subject image part
2 (SP12), and the detection areas ARA2, AR
By performing the zooming operation of the image pickup means 1 so that the size of B2 matches the predetermined reference size (RT1), the subject image portion of the size corresponding to the reference size is always displayed on the display screen PIC. To be displayed.

[0007]

The object image candidate small areas ARA1 and ARB1 are integrated and determined as the detection areas ARA2 and ARB2 of the object image section based on the color characteristics of the object image section in the small area AR, and the size thereof is used as a reference size. By performing the zooming operation of the image pickup means 1 so as to coincide with the above, it is possible to easily realize a video signal tracking method and a video camera system by which a video of a subject image portion having the same size can always be displayed by a simpler configuration. it can.

[0008]

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

In FIG. 1, VCS indicates a video camera system as a whole, and the video camera system VCS is a human face recognition device (Autonomous Target Tracker with H) for a human face model as a subject.
uman Recognition, ATT-HR) system. The video camera system VCS receives the imaged light LA coming from a human face model, which is a subject, through a lens 2 of a lens block unit 1 and an iris 3 in order and received by an image pickup device 4 which is, for example, a CCD (charge coupled device). The image pickup output signal S1 representing the above is given to the signal separation / automatic gain adjustment circuit unit 5.

The signal separation / automatic gain adjustment circuit section 5 samples and holds the image pickup output signal S1, and the image pickup output signal S2 has a predetermined gain by a control signal from an auto iris (AE) system (not shown). The image pickup output signal S2 thus obtained is gain-controlled, and is supplied to the digital camera processing circuit 7 through the analog / digital conversion circuit 6. The digital camera processing circuit 7 forms a luminance signal Y, a chroma signal C and color difference signals R-Y and B-Y based on the image pickup output signal S2, and passes the luminance signal Y and the chroma signal C through a digital / analog conversion circuit 8. And sends out as a video signal S3.

In addition to this, the digital camera processing circuit 7
Supplies the luminance signal Y and the color difference signals RY and BY as the object zoom detection signal S4 to the tracking control circuit section 11, and the tracking control circuit section 11 determines the lens block section based on the object zoom detection signal S4. A follow-up control signal S5 for the zooming drive motor 12 provided for No. 1 is generated.

The follow-up control circuit 11 is a luminance signal Y and a color difference signal RY obtained from the digital camera processing circuit 7.
And BY are applied to the screen reduction circuit 14 via the low-pass filter 13. The screen reduction circuit 14 receives the luminance signal Y of each pixel and the color difference signals RY and BY from the digital camera processing circuit 7 according to the screen reduction control signal S21 sent from the follow-up signal processing circuit 15. As shown in FIG. 3, the display screen PIC is set to 32 × 32 small areas AR or 16
A process of dividing into 16 small areas AR is executed.

Thus, the screen reduction circuit 14 supplies the saturation / hue detection circuit 16 with the color difference signals RY and BY corresponding to 32 × 32 small areas AR or 16 × 16 small areas AR. The hue signal HUE and the saturation signal SAT are supplied to the image memory 17, and the luminance signal Y is directly supplied to the image memory 17. In this case, the image memory 17 is 32 ×
It has a memory area for 32 blocks, and thus pixel data for one field is transmitted from the screen reduction circuit 14 to 32 × 32 small areas or 16 × 16 small areas corresponding to 32 × 32.
Image information for 32 or 16 × 16 blocks is stored in the image memory 17 after being compressed, and as a result, it is generally necessary to prepare a memory capacity for one field pixel, whereas in FIG. In some cases, this can be further simplified.

The saturation / hue detection circuit 16 uses the color difference signal R-
The hue signal HUE and the saturation signal SAT are formed by transforming Y and B-Y into rectangular coordinates / polar coordinates, and thus the human face model as a human subject is represented by the luminance signal Y, the hue signal HUE, and the saturation signal SAT. The subject is recognized based on the visual stimulus that can be perceived. Generally, a visual stimulus that can be perceived by humans is represented by a color coordinate system called an HLS system having an L axis and an SH plane orthogonal to the L axis as shown in FIG. L-axis is brightness (Lightnes
s) and corresponds to the luminance signal Y. The SH plane is represented by polar coordinates orthogonal to the L axis. On the SH plane, S represents the saturation, which is represented by the distance from the L axis. Further, H represents a hue (Hue), and the color difference signal B
The hue is represented by an angle when the −Y direction is 0 °.

In the HLS system solid, when the light source becomes brighter, the color coordinates, that is, the SH plane goes upward along the L axis, and all the colors become white. At this time, the saturation S decreases. On the other hand, when the light source becomes dark, the color coordinates, that is, the SH plane goes downward along the L axis, and all the colors become black. At this time, the saturation S also decreases at the same time. Based on such characteristics of the HLS color coordinate system,
The saturation S and the brightness Y are easily affected by the brightness of the light source,
Therefore, it is difficult to say that the parameter representing the feature amount of the subject is optimal. On the contrary, it can be seen that the hue H is a quantity that is not easily influenced by the light source, as it expresses the characteristic quantity unique to the subject.

However, at any time, when the color of the subject is near the L axis, that is, when the color is white, black, or gray, the signal of the hue H has no meaning as information. It can be seen that, in the worst case, an image with poor S / N may have various hue H vectors even though the image is white. The tracking control circuit unit 11 extracts the characteristics of the human face model, which is the subject, by utilizing the property of the HLS color coordinate system, and drives the zooming drive motor 12 so as to follow the characteristics when the characteristics change. As a result, as the video signal S3, a video signal that follows the movement of the subject and is zoomed is obtained.

That is, the subject zoom detection signal S4 stored in the image memory 17 is given to the address generation circuit 18 by the block designation signal S22 sent from the follow-up signal processing circuit 15 of the microprocessor configuration.
As shown in FIG. 3, an address signal S23 that substantially divides the display screen PIC formed in the image memory 17 into small areas AR having a predetermined size based on the xy rectangular coordinates (x, y) is used. It is supplied to the image memory 17. Thus, the data of each block forming the display screen PIC of the image memory 17 is read out for each small area AR and
Each is evaluated as one image information.

In the case of this embodiment, the display screen PIC has 32 (or 16) small areas AR in the x and y directions, respectively.
And thus the coordinates x = i, y = of the Cartesian coordinates (x, y) for the 32 × 32 (or 16 × 16) small areas AR.
By specifying j, the image information I (x = i, y = j) of the specified small area AR can be read. In the image information I (x = i, y = j) read out from the image memory 17 for each small area AR in this manner, the hue signal HUE component passes through the gate circuit 19 and the skin color detection circuit 20.
On the other hand, the luminance signal Y component is directly applied to the hair detection circuit 21.

The skin color detection circuit 20 detects the image portion of the skin of the human face model. When the hue signal HUE component of each incoming block is within the predetermined skin color range, the skin color candidate detection output S24 is output. Send out. Actually, the dynamic range of the hue signal HUE component is 0 to 359 ° in angle.
Takes a value in the range. On the other hand, the hair detection circuit 21 detects the hair area in the image portion of the human face model. In this embodiment, the dynamic range of the luminance signal Y is represented by the numerical value 0 to 255 (8 bits). Therefore, when the luminance signal Y of each pixel is 50 or less, the pixel is determined to be black, it is determined that the pixel is in the hair region, and the hair candidate detection signal S25 is sent to the tracking signal processing circuit 15.

In the case of this embodiment, the hue noise gate signal forming circuit 2 having a comparator configuration with respect to the gate circuit 19 is used.
5 is provided to compare the saturation signal SAT read from the image memory 17 for each block with the noise determination signal S26 sent from the tracking signal processing circuit 15. When the saturation signal SAT is below a predetermined level, the gate circuit 19 By supplying the gate signal S27 for closing the signal to the gate circuit 19, the hue signal HUE component of the pixel is not input to the skin color detection circuit 20.

Incidentally, when the hue signal HUE detected by the saturation / hue detection circuit 16 is in the vicinity of the L axis (FIG. 2), this is buried in noise because the hue signal HUE has a low saturation. Since there is a possibility that it does not have meaning as information, the hue signal HUE having no meaning is removed in the gate circuit 19. In the above configuration, the follow-up signal processing circuit 15 executes the zoom processing procedure RT1 shown in FIG. 4 so that the face image of almost the same size is always displayed on the display screen PIC.

That is, when the follow-up signal processing circuit 15 enters the zoom processing procedure RT1, it waits for the user to select the face up shot, bust shot or full body full shot at step SP1, and the frame number FN at this time is FN = 1. Therefore, the tracking signal processing circuit 15 initializes the video camera system VCS. After that, the follow-up signal processing circuit 15 moves to step SP2 to extract the face area according to the face extraction processing procedure RT2, and increments the frame number FN to FN + 1.

The face extraction processing procedure RT2 is, as shown in FIGS. 5 and 6, a luminance signal Y, a hue signal HUE, and a saturation signal SA corresponding to each block loaded in the image memory 17.
The feature of the image on the display screen PIC is judged from the skin color detection signal S24 and the hair candidate detection signal S25 obtained based on T, and the size of the face model of the person displayed on the display screen PIC is almost the same. Follow-up control signal S5
To the zooming drive motor 12.

That is, when the follow-up signal processing circuit 15 enters the face extraction processing procedure RT2 of FIG. 5, first, at step SP11.
After dividing the display screen PIC into 32 × 32 small areas AR, in the block of pixels included in each small area AR, in the skin color detecting circuit 20 and the hair detecting circuit 21, the image portion of the block of the pixel is Whether or not the hair is detected is detected, and the skin color candidate detection signal S24 is acquired when the small area is a skin color area candidate, and the hair candidate detection signal S25 is acquired when the small area AR is a hair area candidate.

When the processing of step SP11 is completed in this way, the follower signal processing circuit 15 indicates in the hair candidate small area ARA and the skin color candidate small area A as shown in FIG. 3 or FIG.
The distribution state of the small area AR determined as RB is obtained. Here, as shown in FIG. 3, the hair candidate small areas ARA and the skin color small areas ARB are adjacent to each other without any space, and there is no space between the hair candidate small areas ARA and the skin color candidate small areas ARB. If such a candidate extraction result is obtained, it means that in this case, information that matches the actual condition of the human face model as the subject has been obtained.

However, since the direction of the light source with respect to the subject and the manner of reflection are not uniform in general, as shown in FIG. 7, the hair candidate small area ARA and the skin color candidate small area A are shown in FIG.
A blank area is formed in the collection of RBs, and in addition to this, a collection of hair candidate small areas ARB and a skin color candidate small area ARB.
It is often possible to obtain an extraction result in which a small blank area is formed even between groups of.

The "merging rule 1" of the discontinuous area is "IF I
(I + p, j + q) = hair region candidate, (p, q = + 1)
or-1or0, except p = q = 0) Then I
(I, j) = hair region candidate ”. The “merging rule 1” of this discontinuous area is the address (x = i, y
= J) area, if there is a hair candidate small area ARA in a small area adjacent to the small area AR, the image information I (i, i, i) of the specified small area (x = i, y = j) is obtained.
j) is merged into the hair candidate small area ARA.

For example, in FIG. 7, a small blank area (x = 6,
When y = 5) (marked with a triangle) is the designated area,
A small area AR surrounding the blank small area (x = 6, y = 5), that is, (x = 5, y = 4) (x = 6, y = 4)
(X = 7, y = 4) (x = 7, y = 5) (x = 7, y =
6) (x = 6, y = 6) (x = 5, y = 6) and (x =
5, y = 6) small area AR has a hair candidate small area ARA
Since the hair area (x = 6, y = 4) is included as 1, the blank small area (x =
6, y = 5) is merged into a group of hair candidate small areas ARA1. The "merge rule 2" of the discontinuous area is applied after the "merge rule 1" is applied, and the content is "IF.
I (i, j) = skin color region candidate andI (i + p, j +
q) = skin color region candidate (p, q = 1 or-1 or 0, except p = q = 0) Then I (i, j) = skin color region candidate ”.

The "merging rule 2" of this discontinuous area is similar to that described above for the hair candidate small area ARA, and for the designated small area (x = i, y = j), there is a skin color area candidate in the area adjacent thereto. For example, the designated small area (x =
The image information I (x = i, y = j) of i, y = j) is merged as the skin color candidate small area ARB. For example, in FIG. 7, when the blank small area (x = 6, y = 6) (marked with an X) is designated as the designated small area, the blank small area (x =
6, y = 6) surrounding small area AR, that is, (x =
5, y = 5) (x = 6, y = 5) (x = 7, y = 5)
(X = 7, y = 6) (x = 7, y = 7) (x = 6, y =
7) The small area AR of (x = 5, y = 7) and (x = 5, y = 6) includes the small skin area (x = 7, y = 7) as the small skin area ARB1. Therefore, according to the “merging rule 2”, the blank small area (x = 6, y = 6) is merged into the set of skin color candidate small areas ARB1.

When the merge processing of the discontinuous areas of step SP12 is completed in this way, as shown in FIG. 7, there is a blank small area between the collection of hair candidate small areas ARA1 and therefore the collection of hair candidate small areas ARA1 is not complete. What is continuous can be filled with the merge area ARA2 as shown in FIG. Similarly for the skin color candidate small area ARB1, as shown in FIG. 7, since there is a space between the group of skin color candidate small areas ARB1, the skin color candidate small area AR is present.
When B1 is discontinuous, the blank can be filled with the merged skin color candidate small area ARB2 as shown in FIG.

The follow-up signal processing circuit 15 has a step SP12.
When the processing of (1) is completed, the process proceeds to step SP13 to determine the hair region and the skin color region by the "hair region determination rule 1" and the "skin color region determination rule 2". “Hair area determination rule 1” is “IF I (i,
j) = hair region candidate and I (i + p, j + q) = hair region candidate (p, q = + 1 or-1 or 0, where p = q
= 0 except) Then I (i, j), I (i + p, j +
q) as the hair area ”.

This "hair region determination rule 1" is applied when a hair candidate small region is located in a small region adjacent to the small region of the address (x = i, y = j) with the small region as the center. This means that the image information I (i, j) of the designated area (x = i, y = j) is integrated into the small hair area group HR. In addition, "Skin color area determination rule 2" is "IF
I (i, j) = skin color region candidate and I (i + p, j +
q) = skin color region candidate (p, q = + 1or-1or0, except p = q = 0) Then I (i, j), I (i +
p, j + q) as the skin color region ”.

This "skin color region determination rule 2" is similar to that described above for the hair candidate small region, and for the designated small region (x = i, y = j), there is a skin color candidate small region adjacent to this. For example, the designated small area (x = i,
The image information I (x = i, y = j) of y = j) and the image information of the adjacent area having the skin color candidate small area are integrated into the skin color small area group FC.

Then, for example, as described above, the data is merged into the small hair region according to "merging rule 1 of the hair region".
By applying the “rule of hair region determination rule 1” to the marked area (x = 6, y = 5) (FIG. 7), the small area around the small area (x = 6, y = 5), that is, (X = 5,
y = 4) (x = 6, y = 4) (x = 7, y = 4) (x =
7, y = 5) (x = 7, y = 6) (x = 6, y = 6)
In the small areas of (x = 5, y = 6) and (x = 5, y = 5), small hair areas (x = 5, y =
5), (x = 5, y = 4), (x = 6, y = 4) (x =
7, y = 4) and (x = 7, y = 5) are included, the region (x = 5, y = 4) (x = 6, y = 4) (x =
7, y = 4) (x = 7, y = 5) (x = 6, y = 5) and (x = 5, y = 5) are integrated as a hair sub-region group FC1.

Further, for example, as described above, for the small area (x = 6, y = 6) marked with an X, which is merged into the small skin color area according to the "skin color area merging rule 2", "skin color area determination" is performed. Rule 2 ”is applied, a small area (x = 6, y =
6) around the small area, that is, (x = 5, y = 5) (x
= 6, y = 5) (x = 7, y = 5) (x = 7, y = 6)
(X = 7, y = 7) (x = 6, y = 7) (x = 5, y =
7) and (x = 5, y = 6) small areas as skin color area candidates (x = 7, y = 6), (x = 7, y =
7) and (x = 6, y = 7) are included, the region (x = 6, y = 6) (x = 7, y = 6) (x = 7, y =
7) and (x = 6, y = 7) are integrated as the skin color small area group FC1.

In this way, steps SP12 and SP
As shown in FIG. 8, the display screen PIC displays the small area groups HR1 and HR by merging the discontinuous small areas and integrating the small hair area and the skin color small area in FIG.
2 is integrated as a hair area, and the small area groups FC1 and FC2 are integrated as a skin color area. After that, the tracking signal processing circuit 15 moves to step SP14 of FIG. 6 to integrate the hair regions HR1 and HR2 of the hair and the skin color region FC.
Area of 1 and FC2 Area Hair and Area Calculate the Skin. In the case of the display screen PIC of FIG. 8, the hair area HR1
Area of Hair HR1 is Area as the total number of areas included in this Hair HR1 = 45 is calculated, and in the same way, the area of the hair region group HR2 is Area. Hair HR2 = 1
4. Area of skin color area group FC1 is Area Skin FC1 = 6
6. Area of skin color area group FC2 is Area Skin FC2 = 1
It is also calculated as 0.

Then, the follow-up signal processing circuit 15 executes step S.
The process proceeds to P15 to determine the face area on condition that the conditions of "face area determination rule 1" and "face area determination rule 2" are satisfied. "Face region determination rule 1" is 1
Considering one hair small area group and one skin color small area group as one area, the area Area of the one hair area group Hair
And area of one skin color small area group Area The area ratio with Skin is
`` IF (Area Hair) / (Area Skin) <5and (Area
Hair) / (Area Skin)> 1/5 Then it is a face area candidate ”.

According to the "face determination rule 1", if the area ratio of the hair small area group and the skin color small area group is within 5 times and 1/5 times or more, the set of the small area groups can be the face area. Determined to be It is the display screen PI that makes such a determination.
In an image in which there are a large number of small areas with a dark brightness level on C (for example, when there is a dark curtain in the background), all of these dark small areas may be judged as hair areas, so this is prevented, and the same applies. If there are a large number of small flesh-colored areas other than the flesh-colored area of the face, it may not be possible to correctly determine the face area, and this erroneous determination is prevented.

The "face area determination rule 2" is "IF I
(I, j) = hair area and I (i, j-1) = skin color area Then Neighbor = Neighbor + 1 IF Neighbor>
3 Then the hair area + skin color area is the face. " In this "face area determination rule 2", the upper side of the screen is the small hair area group (that is, I (i, j)), and the lower side is the skin color small area group (that is, I (i, j-).
1)), and if there is a condition that the hair small area group and the skin color small area group are in contact with each other in at least three small areas, the set of the hair small area group and the skin color small area group is formed on the face. Judge that there is.

By applying this "face area determination rule 2", a pair of small area groups, that is, a hair small area group HR1 and a skin color small area group F in the display screen PIC of FIG.
Since the set C1 is in contact with eight small areas, it satisfies the rule 2, and thus it is determined that the face is one face. On the other hand, since there is no small area in contact with the pair of area groups, that is, the set of the hair small area group HR2 and the skin color small area group FC2, the condition of Rule 2 is not satisfied, and accordingly, the face area is determined. Excluded.

Thus, the follow-up signal processing circuit 15 finishes the face extraction processing, and returns from the step SP16 to the main routine of FIG. Then, the tracking signal processing circuit 15 moves to step SP3 in the main routine of FIG.
Area of the face area determined in P2 Area Find Face as "hair area + skin color area". Thus, the follow-up signal processing circuit 15 detects the area of the face area for the frame number FN = 1, and then moves to step SP4, where the reference area Area of the face corresponding to the currently selected type of shot area. St
D is drawn from the face area table.

Here, the face area table is a reference area Area when the upshot mode is selected for two types of screen division methods, that is, 32 × 32 blocks and 16 × 16 blocks, respectively. Std is set to 400 and 100, and the standard area Area when the bust shot is selected
Standard area Area when Strud = 100 and 25 is selected, and Furushoto is selected Std has been selected for 25 and 7. Subsequently, the follow-up signal processing circuit 15 proceeds to step SP5 to compare the reference face area with the current detected face area of the subject, and supplies the zoom drive motor 12 with the follow-up control signal S5 such that the difference becomes zero.

At this time, the control amount Co by the follow-up control signal S5
ntrol Signal is rule 1 and
And rule 2 are decided. Rule 1 is for screen
If the ratio is 16x16 block, "IF screen division = 16x16
 Then Control Signal = (Area Face-Area Std)
, The detected face area Area as the controlled variable Control Signal
Face and standard face area Area Value that is 4 times the value of the difference from Std ”
Is required as. On the other hand, rule 2 is 32 × 32 block
It specifies the control amount when the screen is divided into
"IF screen division = 32 x 32 Then Control Signal = (Area
Face-Area Std) and the controlled variable Control Signal
Detected face area Area Face and standard face area Area With Std
Difference value ”.

Thus, the follow-up signal processing circuit 15 finishes the zoom process for the frame number FN = 1, and then returns to the above-mentioned step SP2 and the next frame FN = FN +.
The zoom process for 1 is repeatedly executed. When the tracking signal processing circuit 15 repeats such zoom processing, the zooming drive motor 12 causes the zooming drive motor 12 to subtract the current detected face area value of the lens 2 of the lens block unit 1 from the face area table at step SP4. The zoom is controlled to a value almost the same as Std, and thus the size of the face image displayed on the display screen PIC is always the reference face area Area. It will maintain the size equivalent to Std.

With the above arrangement, even when the face model as the subject approaches the camera or moves away from the camera, the face image of the same size is always displayed on the display screen PIC. The lens 2 of the lens block can be zoom-controlled so that it can be taken out, thus eliminating the need to provide means such as a distance sensor as used in single-lens reflex cameras,
Therefore, it is possible to realize a video camera system having a much simpler configuration as a whole.

In the above-described embodiment, the case where the lens 2 of the lens block unit 1 is optically zoomed by the zooming drive motor 12 as the zooming means has been described. Not limited to this, for example, a digital video signal obtained based on the image pickup signal S1 obtained from the solid-state image pickup device 4 of the lens block unit 1 is taken into a frame memory, and the frame memory is converted into a follow-up control signal S5 obtained from the follow-up signal processing circuit 15. Even if the zoom process is performed by obtaining an image that is actually enlarged or reduced by reading by the corresponding reading method, the same effect as in the above case can be obtained.

[0047]

As described above, according to the present invention, the subject image candidate small areas are integrated and determined as the detection area of the subject image portion based on the color characteristics of the subject image portion in the small area.
By performing the zooming operation of the image pickup means so that the size thereof matches the reference size, a video signal tracking method and a video that can always display a video image of a subject image part having the same size with a simpler configuration. A camera system can be easily realized.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a video camera system according to the present invention.

FIG. 2 is a schematic diagram for explaining an HLS color coordinate system that represents a visual stimulus.

FIG. 3 is a schematic diagram showing a display screen PIC in which image information is obtained by dividing it into small areas.

FIG. 4 is a flow chart showing a zoom processing procedure.

FIG. 5 is a flow chart showing a face extraction processing procedure.

FIG. 6 is a flowchart showing a face extraction processing procedure following FIG. 5;

FIG. 7 is a schematic diagram showing a display screen PIC having a detected hair region and skin color region.

8 is a schematic diagram showing a display screen PIC obtained by merging and integrating the hair region and the skin color region of FIG. 7.

[Explanation of symbols]

VCS ... video camera system, 1 ... lens block section, 2 ... lens, 3 ... iris, 4 ... solid-state image sensor, 5 ... signal separation / automatic gain adjustment circuit section, 7 ... digital camera processing circuit , 11 ... Tracking control circuit section, 1
2 ... Zooming drive motor, 14 ... Screen reduction circuit,
15 ... Follow-up signal processing circuit, 16 ... Saturation / hue detection circuit, 17 ... Image memory, 18 ... Address generation circuit, 20 ... Skin color detection circuit, 21 ... Hair detection circuit.

Continuation of the front page (56) Reference JP-A-2-87876 (JP, A) JP-A-3-77927 (JP, A) JP-A-63-223974 (JP, A) JP-A-4-60769 (JP , A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N 5/222-5/247

Claims (2)

(57) [Claims] 1. A video signal tracking method for forming an output video signal for displaying an image including a subject image portion on a display screen based on an image pickup signal obtained from an image pickup means which performs a zooming operation, which is displayed on the display screen. The image to be divided is divided into a plurality of small areas, and based on the color characteristics of the image pickup signal corresponding to the small areas.
Te extracts object image candidate small area serving as candidates of the subject image portion, the blank subregion during collection of the subject image candidate subregion
When it occurs, the subject image candidate in the blank small area
The blank small area adjacent to the small area is the subject image candidate.
The size of the detection area is determined as a detection area of the subject image portion in combination with a small area, and the size of the detection area is a predetermined reference size.
By performing the zooming operation of the image pickup means so as to coincide with the above, the size of the reference is constantly displayed on the display screen.
A method for tracking a video signal, characterized in that the subject image portion having a size corresponding to is displayed . 2. A video camera system for forming an output video signal for displaying an image including a subject image portion on a display screen based on an image pickup signal obtained from an image pickup means which performs a zooming operation, which is displayed on the display screen. means for dividing said image into a plurality of small areas to be, based on the color characteristics of the image pickup signal corresponding to the small region
Means for extracting a subject image candidate small area that is a candidate for the subject image section, and a blank small area between the group of subject image candidate small areas.
When it occurs, the subject image candidate in the blank small area
The blank small area adjacent to the small area is the subject image candidate.
Means for merging with a small area to determine as a detection area of the subject image portion, and a reference size for which the size of the detection area is predetermined
By performing the zooming operation of the image pickup means so as to coincide with the above, the size of the reference is constantly displayed on the display screen.
For displaying the subject image portion having a size corresponding to
Video camera system which is characterized in that it comprises and.