KR101545408B1 - Method for detecting profile line and device for detecting profile line - Google Patents

Abstract

A method for detecting a side contour is disclosed. According to an embodiment of the present invention, there is provided a detection method including the steps of: acquiring a first image including a side contour; detecting a first area based on a predetermined facial recognition method in the first image; Detecting a second region based on the branch point information, extracting a color of the second region, and detecting a first side image from the first region based on the color of the extracted second region and And detecting a side contour from the first image based on the first side image. In addition, the detection method may include detecting a reference point located at an outermost edge of the first side image, moving a searcher from the first image in the first area to the left and right with reference to the reference point, . In the detection method, contour point detection can be repeated in the vertical direction to detect the side contour line.

Description

TECHNICAL FIELD [0001] The present invention relates to a lateral contour detection method and a lateral contour detection apparatus,

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a side contour detection method and a side contour detection apparatus, and more particularly, to a side contour detection method and a side contour detection apparatus that detect a side contour from a facial side photographed image.

In recent years, the development of methods for facial recognition has been actively under way. A conventional facial recognition method captures a facial image including a face of a subject. In addition, the conventional facial recognition method can extract facial features from a captured facial image, and then perform facial recognition based on the extracted feature point information.

For example, a conventional facial recognition method can extract facial parts such as eyes, nose, and mouth based on feature points extracted from a facial image. In addition, the conventional face recognition method can detect the front face of the face based on the extracted face portion. More specifically, the conventional facial recognition method can compare the stored feature point information with the detected front face to determine who is the person corresponding to the image.

However, the conventional facial recognition method relates to a method of determining a person corresponding to an image by simply grasping a feature point, and does not accurately detect the shape of the face. In addition, a method of detecting a side image of a face has not been disclosed in the prior art, and a method of detecting a side image, which is one of the person characteristics, is required.

In particular, it is possible to grasp information related to the user's health based on the contour of the side. For example, it is possible to grasp the health information of the user based on the shape or size of the forehead, nose, lips, etc., and accordingly, development of a method of detecting a side image including a contour of a side is required.

This disclosure, which is devised in response to the above-described technology development request, provides a method and apparatus for detecting a side contour of a face from a facial image.

A side contour detection method according to an embodiment includes the steps of acquiring a first image including a side contour, detecting a first region based on a predetermined facial recognition method in the first image, Detecting a second area based on predetermined branch point information, extracting a color of the second area, and detecting a first side image from the first area based on the extracted color of the second area And detecting a side contour from the first image based on the first side image.

The side contour detection method may further include detecting a reference point existing at an outermost edge of the first side image.

The method of detecting a contour of a side includes detecting an outline in a vertical direction from a first image in a first area on the basis of the detected reference point; And detecting the side contour based on the detected contour.

The detecting of the contour may include: setting a search interval of a predetermined distance extending in the left-right direction around the detected reference point; Calculating a texture difference between a left region of the searcher and a right region of the searcher while moving the searcher in the set search interval; Determining a point at which the calculated difference is maximum as a contour point constituting the contour; And detecting the outline by repeating the detection of the outline point in the up and down direction based on the detected reference point.

The calculating of the texture difference between the left region of the searcher and the right region of the searcher may include calculating at least one of a histogram of the left region of the searcher and a correlation function, a Chi-square value, and a Hellinger distance value of the histogram of the right region of the searcher. Generating one; And computing the difference based on at least one of the correlation function, the Chi-square value, and the Hellinger distance value.

Here, the step of detecting the first area may detect the rectangular area including the side contour as the first area based on the adaboost method or the lateral face detection method based on the skin color information.

The step of detecting the second area may detect the second area of a polygonal shape based on a predetermined branch point of the rectangular area.

In addition, the bifurcation point may be a presentation point, an opian, a jaw node, and a crossover point, and the presentation point, the opian, the jaw node, and the crossover point may be previously set by relative average position information.

In addition, the step of detecting the first side image from the first area may detect, as the first side image, an area in the first area where the degree of similarity with the color of the extracted second area exceeds a threshold value .

The detecting the reference point existing at the outermost edge of the first side image may include detecting a third region from the second region based on preset position information; And detecting, as the reference point, a point existing at an outermost edge of the first side image in the third area.

According to another aspect of the present invention, there is provided an apparatus for detecting a contour of a side surface, comprising: an image obtaining unit obtaining a first image including a side contour; And detecting a first region based on the facial recognition method predetermined in the first image, detecting a second region based on the predetermined branch point information in the first region, extracting a color of the second region, And a processor for detecting a first side image from the first region based on the extracted color of the second region and detecting a side contour from the first image based on the first side image.

The processor detects a reference point existing at an outermost edge of the first side image.

The processor can detect a vertical contour line from the first image in the first area on the basis of the detected reference point and detect the side contour line on the basis of the detected contour line.

The processor sets a search area of a predetermined distance extending in the left and right direction around the detected reference point and moves the searcher on the search area to calculate a texture difference between the left area of the searcher and the right area of the searcher , A point at which the calculated difference is maximum is determined as a contour point constituting the contour line, and the contour line can be detected by repeating contour point detection in the up and down direction based on the detected reference point.

The processor generates at least one of a correlation function, a Chi-square value, and a Hellinger distance value for a histogram of a left area of the searcher and a histogram of a right area of the searcher, and calculates at least one of a correlation function, a Chi- Value based on at least one of the values.

The processor may detect a rectangular area including the side contour as the first area based on an adaboost method or a side face detection method based on skin color information.

The processor can detect the second area in the form of a polygon based on a predetermined branch point of the rectangular area.

Here, the bifurcation point may be a presentation point, an opian, a jaw node, and a crossover point, and the presentation point, the opian, the jaw node, and the crossover point may be set by relative average position information.

The processor may detect, as the first side image, an area in the first area where the degree of similarity with the color of the extracted second area is greater than a threshold value.

The processor can detect a third region from the second region based on preset position information and detect a point existing at an outermost portion of the first side image in the third region as the reference point.

According to various embodiments of the present disclosure, a method and apparatus for detecting a side contour of a face from a facial image can be provided. Accordingly, the side contour line can be managed as a foreign material characteristic.

1 is a block diagram of a side contour detection apparatus according to an embodiment.
FIG. 2 is a flowchart for explaining a side contour detection method according to an embodiment.
3A is a conceptual diagram of an input image.
3B is a conceptual diagram for explaining the first area.
4 is a conceptual diagram for explaining the second area.
5A and 5B are conceptual diagrams for explaining the first side image detection.
6 is a flowchart illustrating a method of detecting a side contour according to an embodiment of the present invention.
7 is a conceptual diagram for explaining the third area.
8A and 8B are conceptual diagrams for explaining contour detection.
FIG. 9 is a conceptual diagram for explaining an outline according to an embodiment.

In the following, some embodiments will be described in detail with reference to the accompanying drawings. However, it is not limited or limited by these embodiments. Like reference symbols in the drawings denote like elements.

Although the terms used in the following description have selected the general terms that are widely used in the present invention while considering the functions of the present invention, they may vary depending on the intention or custom of the artisan, the emergence of new technology, and the like.

Also, in certain cases, there may be terms chosen arbitrarily by the applicant for the sake of understanding and / or convenience of explanation, and in this case the meaning of the detailed description in the corresponding description section. Therefore, the term used in the following description should be understood based on the meaning of the term, not the name of a simple term, and the contents throughout the specification.

1 is a block diagram of a side contour detection apparatus 100 according to an embodiment. As shown in FIG. 1, the side contour detecting apparatus 100 may include an image obtaining unit 110 and a processor 120.

Here, the image obtaining unit 110 may obtain a first image including a side contour line. In one embodiment, the image acquisition unit 110 may be implemented with various imaging devices such as a CCD or a photodiode capable of imaging. When the image acquisition unit 110 is implemented as an imaging device, the still image or the moving image may be captured under the control of the processor 120. [ In addition, the image acquiring unit 110 may include an auxiliary light source (e.g., a flash (not shown)) that provides a light amount required for photographing. The image obtaining unit 110 may be physically independent of and electrically connected to the side contour detecting apparatus 100 such as a digital camera or a camcorder.

In another embodiment, the image acquisition unit 110 may be implemented as an interface for receiving a first image including a side contour line. For example, a first image including a side contour may be stored in a storage unit (not shown). Alternatively, the first image including the side contour may be received from an external source and temporarily stored. The image acquiring unit 110 may acquire a first image by receiving a first image from a storage unit (not shown) or a communication unit (not shown).

The processor 120 detects the first area based on the facial recognition method predetermined in the first image, detects the second area based on the predetermined point information in the first area, and determines the color of the second area And extract a side contour from the second area based on the color of the extracted second area. More detailed operation of the processor 120 will be described later in more detail.

FIG. 2 is a flowchart for explaining a side contour detection method according to an embodiment.

In step 210, the lateral contour detection method may obtain a first image including a side contour. As described above, the side contour detection method may acquire and acquire a first image, or may acquire and acquire a first image from another hardware. For example, the side contour detection method can obtain an input image as shown in FIG. 3A.

In step 220, the lateral contour detection method may detect a first region including a side contour from the first image. For example, the side contour detection method can detect the first area 310 as shown in FIG. 3B. In one embodiment, the first region may be in the form of a rectangle including side contours.

The side contour detection method according to an exemplary embodiment can detect the first area based on, for example, the adaboost method. More specifically, the side contour detection method can detect the first area based on the comparison with the pre-stored pattern.

The side contour detection method according to another embodiment can detect the first area based on, for example, the color detection method. For example, the side contour detection method can store the color distribution data of the side image. The side contour detection method can store color distribution data such as average and deviation of color of a general side image. The side contour detection method can compare previously stored color distribution data among the photographed first images, and can detect the first region based on the comparison result. For example, the side contour detection method can set a rectangular comparison window and move the comparison window on the first image. The side contour detection method can determine the similarity of the color distribution between the comparison window and the first image and can detect the region corresponding to the position of the comparison window having the highest similarity as the first region.

Referring back to FIG. 2, the side contour detection method can detect the second region from the first region based on the predetermined branch point information. For example, as shown in Fig. 4, the lateral contour detection method can detect the branch points 401 to 404 in the first area. The bifurcation point may include an origination point 401, an offlier 402, a jaw node 403, and a point of intersection 404.

Bifurcation points 401 to 404 may be preset for use in samples of various subjects. For example, the lateral contour detection method according to an embodiment may receive the positions of the presentation point, the opian, the jaw node, and the crossover point in each of the first regions of the samples of various subjects. Alternatively, the lateral contour detection method may receive the average position of the presentation point, the opian, the jaw node, and the crossover point. The lateral contour detection method can set the branch points 401 to 404 based on the average position of each of the presentation point, the opian, the jaw node, and the crossover point. That is, the presentation point 401 may be an average point of the presentation point in the first sample region of various subjects. Each of the opiates 402, the jaw node points 403 and the crossover points 404 may also be positions of the opiate, jaw node, and crossover point average in the sample first region of various subjects.

The side contour detection method can detect the second region by connecting the branch points 401 to 404. In FIG. 4, the second region may be a polygonal shape in which the branch points 401 to 404 are connected.

The second region may be a region in which the face color of the subject is mainly distributed. The presentation point, the opiate, the jaw node, and the crossover point may also be preset as areas where the subject's facial color is mainly distributed.

Referring again to FIG. 2, in step 240, the side contour detection method may extract the color of the second area. In the lateral contour detection method in step 250, an area having a degree of similarity with the color of the extracted second area in the first area may be detected as a side contour candidate, e.g., a first contour line.

For example, as in FIGS. 5A and 5B, the lateral contour detection method may extract the color of the second area 505. FIG. The side contour detection method can extract the color distribution of the second area 505. [ For example, the lateral contour detection method can measure the color distribution based on the color coordinate information of each pixel in the second area 505, thereby extracting the color of the second area 505. [

The side contour detection method can detect an area where the degree of similarity with the color of the second area 505 exceeds a threshold value as a first side image. For example, as shown in FIG. 5B, the side contour detection method may detect the first side image 510. The side contour detection method compares the color coordinate information of the second area 505 and the color coordinate information of each pixel of the first area, and classifies it as the first side image 510 when the similarity is over a threshold value. The side contour detection method can detect the sorted pixel set as the first side image 510.

According to the above description, the side contour detection method can detect the side contour candidate, i.e., the first side image.

6 is a flowchart illustrating a method of detecting a side contour according to an embodiment of the present invention.

In step 610, the lateral contour detection method may detect the side candidate region, i.e., the third region from the first side image. The side contour detection method can detect a predetermined position in the first side image as the third area.

As shown in FIG. 7, the side contour detection method can detect the third area 720 from the first side image 710. For example, the third region 720 may be determined based on the relative point-of-interest location information of the side, i.e., the nose end point. Here, the location information of the non-point can also be preset based on the samples of various subjects.

In step 620, the lateral contour detection method may extract a reference point in a third region. As shown in FIG. 7, the side contour detection method can extract the point 721 located at the outermost one of the third regions 720 as a reference point. More specifically, the side contour detection method can extract, as a reference point, a point having a maximum value or a minimum value among coordinate values in the left and right direction of the third region 720.

In step 630, the side contour detection method can detect contour points by moving the searcher in the left and right directions with reference to the reference point. In addition, in step 640, the side contour detection method can detect contour lines by repeating contour point detection in the upper direction.

For example, the side contour detection method can set a searcher for a preset distance in the left and right directions based on the reference point 800 as shown in FIG. 8A. FIG. 8B is an illustration of the searcher 810. FIG. 8B, the searcher 810 may have the sizes of M ₁ and M ₂ on the left and right with respect to the central branch point 811, respectively. In addition, the searcher 810 may have a size of N up and down. On the other hand, M ₁ + M ₂ may be set to M.

The side contour detection method can calculate the texture difference value of the area on the left side and the area on the right side based on the central branch point 811 in the searcher 810. [ For example, the side contour detection method can determine the histogram (H ₁ ) of the left region and the histogram (H ₂ ) of the right region. In addition, the lateral contour detection method can calculate the difference between two histogram shapes. More particularly, the lateral contour detection method can derive a correlation function of two histograms and calculate a texture difference value based on the derived correlation function value. The correlation function according to an embodiment may be expressed by Equation (1).

및

는 수학식 2를 따를 수 있다.In Equation (1)

And

(2) < / RTI >

및

는 히스토그램의 평균을 의미할 수 있다.In other words,

And

May mean the average of the histogram.

The side contour detection method can detect a point having the greatest texture difference in the right and left regions based on the correlation function. The side contour detection method can set a search interval of a preset size and can determine the texture difference at each point while moving the searcher 810 to the left and right within the search interval. Here, the search range may be within the first area 310. [ The side contour detection method can detect the point having the greatest texture difference as the contour point constituting the contour line.

When the contour point detection is completed in one y-axis coordinate, the side contour detection method can detect the contour point in the y-axis coordinate by moving the searcher 810 in the upward direction 821. The side contour detection method can detect an upper contour point. When the detection of the contour point in the upper direction 821 is completed, the side contour detection method can detect the contour point in the y-axis coordinate by moving the searcher 810 in the downward direction 822. The side contour detection method can detect the lower contour point. The side contour detection method can detect the side contour by moving the searcher 810 from the minimum value to the maximum value of the first image in the first area.

According to the above description, the lateral contour detection method can detect the side contour which is a set of contour points.

Meanwhile, according to another embodiment, the side contour detection method can derive the Chi-square values of the two histograms and calculate the texture difference value based on the derived Chi-square values. The side contour detection method can derive a Chi-square value based on Equation (3).

The lateral contour detection method can detect the point where the texture difference of the left and right regions is greatest based on the Chi-square value. The side contour detection method can determine the texture difference at each point while moving the searcher 810 to the left and right. The side contour detection method can detect the point having the greatest texture difference as the contour point constituting the contour line. The side contour detection method can detect the outline by moving the searcher 810 in the up and down directions and then moving the searcher 810 in the left and right directions again from the respective y coordinates.

Meanwhile, according to another embodiment, the side contour detection method can derive the Hellinger distance value of two histograms and calculate the texture difference value based on the derived Hellinger distance value. The side contour detection method can derive a Chi-square value based on Equation (4).

The lateral contour detection method can detect the point where the texture difference of the right and left regions is greatest based on the Hellinger distance value. The side contour detection method can determine the texture difference at each point while moving the searcher 810 to the left and right. The side contour detection method can detect the point having the greatest texture difference as the contour point constituting the contour line. The side contour detection method can detect the outline by moving the searcher 810 in the up and down directions and then moving the searcher 810 in the left and right directions again from the respective y coordinates.

The search range may be set in the vertical direction as well, and the search range in the vertical direction may also be set in the first image in the first area 310.

9 is a conceptual diagram of a detected side contour line 910 by moving a searcher starting from a reference point 900 according to an embodiment. As shown in FIG. 9, the side contour 910 can be detected as a set of contour points within the first region 310.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA) A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions.

The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software.

For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded.

The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software.

Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like.

Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like.

The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the equivalents of the claims, as well as the claims.

Claims (20)

Obtaining a first image including a side contour;
Detecting a first region based on a face recognition method predetermined in the first image;
Detecting a second area based on predetermined branch point information in the first area;
Extracting a color of the second area;
Detecting a first side image from the first region based on the color of the extracted second region;
Detecting a reference point existing at an outermost edge of the detected first side image;
Detecting a vertical contour line from the first image in the first area on the basis of the detected reference point; And
Detecting a side contour line based on the detected contour line in the up and down direction
Wherein the side contour detection method comprises: delete delete The method according to claim 1,
The step of detecting the up-down contour includes:
Setting a search range of a predetermined distance extending in the left-right direction around the detected reference point;
Calculating a texture difference between a left region of the searcher and a right region of the searcher while moving the searcher in the set search interval;
Determining a point at which the calculated difference is maximum as an outline point constituting the outline of the up and down direction; And
Repeating the detection of the contour points in the vertical direction on the basis of the detected reference points and detecting the contour lines in the up and down directions
Wherein the side contour detection method comprises: 5. The method of claim 4,
Wherein the step of calculating a texture difference between a left region of the searcher and a right region of the searcher comprises:
Generating a correlation function, a Chi-square value, and a Hellinger distance value for a histogram of a left region of the searcher and a histogram of a right region of the searcher; And
Calculating the difference based on at least one of the correlation function, Chi-square value and Hellinger distance value
Wherein the side contour detection method comprises: The method according to claim 1,
Wherein the detecting the first region comprises:
and detecting a rectangular area including the side contour as the first area based on a side face detection method based on an adaboost method or skin color information. The method according to claim 6,
Wherein the detecting the second region comprises:
And detecting the second region in the form of a polygon based on a predetermined branch point of the rectangular region. 8. The method of claim 7,
Wherein the bifurcation point is an origination point, an opian, a jaw node, and a crossover point, and the presentation point, the opponent, the jaw node, and the crossover point are preset by relative average position information. 8. The method of claim 7,
Wherein detecting the first side image from the first region comprises:
Wherein the first side image is an area in which the degree of similarity with the color of the extracted second area is greater than a threshold value in the first area. Obtaining a first image including a side contour;
Detecting a first region based on a face recognition method predetermined in the first image;
Detecting a second area based on predetermined branch point information in the first area;
Extracting a color of the second area;
Detecting a first side image from the first region based on the color of the extracted second region;
Detecting a side contour from the first image based on the first side image;
Detecting a third region from the second region based on predetermined positional information; And
Detecting as a reference point a point existing at an outermost edge of the first side image in the third area
Wherein the side contour detection method comprises: An image acquiring unit acquiring a first image including a side contour; And
Detecting a first region on the basis of a face image recognition method predetermined in the first image, detecting a second region based on predetermined branch point information in the first region, extracting a color of the second region, A processor for detecting a first side image from the first region based on the extracted color of the second region and for detecting a side contour from the first image based on the first side image,
Lt; / RTI >
Wherein the processor detects a reference point existing at an outermost edge of the first side image and detects a contour line in a vertical direction from the first image in the first area on the basis of the detected reference point, A side contour detection apparatus for detecting the side contour line based on contour lines. delete delete 12. The method of claim 11,
The processor sets a search area of a predetermined distance extending in the left and right direction around the detected reference point and moves the searcher on the search area to calculate a texture difference between the left area of the searcher and the right area of the searcher Determining a point at which the calculated difference is maximum as a contour point constituting the contour line and detecting the contour point by repeating contour point detection in a vertical direction with reference to the detected reference point. 15. The method of claim 14,
The processor generates at least one of a correlation function, a Chi-square value, and a Hellinger distance value for a histogram of a left area of the searcher and a histogram of a right area of the searcher, and calculates at least one of a correlation function, a Chi- And calculating the difference based on at least one of the values. 12. The method of claim 11,
Wherein the processor detects a rectangular area including the side contour as the first area based on an adaboost method or a facial face detection method based on skin color information. 17. The method of claim 16,
Wherein the processor detects the second region in the form of a polygon based on a predetermined branch point of the rectangular region. 18. The method of claim 17,
Wherein the bifurcation point is a presentation point, an opian, a jaw node, and a crossover point, and the presentation point, the opponent, the jaw node, and the crossover point are preset by relative average position information. 18. The method of claim 17,
Wherein the processor detects, as the first side image, an area in the first area that is similar to the color of the extracted second area by more than a threshold value. An image acquiring unit acquiring a first image including a side contour; And
Detecting a first region on the basis of a face image recognition method predetermined in the first image, detecting a second region based on predetermined branch point information in the first region, extracting a color of the second region, A processor for detecting a first side image from the first region based on the extracted color of the second region and for detecting a side contour from the first image based on the first side image,
Lt; / RTI >
Wherein the processor detects a reference point existing at an outermost edge of the first side image and detects a third area from the second area based on preset position information and detects an outermost edge of the first side image in the third area, As the reference point. ≪ Desc / Clms Page number 13 >

Images