KR20160124574A - Face recognition system and method - Google Patents

Abstract

A face recognition method according to the present invention includes the steps of receiving an input image, extracting facial contour data from the input image, calculating feature elements of the input image using the facial contour data, Calculating a similarity value of each of the candidate IDs selected according to a result of the comparing, comparing the calculated value of each of the candidate IDs with the stored values of the corresponding feature elements stored in advance, And outputting at least one similar ID out of the candidate IDs according to a degree of similarity between the similarity value and a predetermined threshold value, wherein the feature elements of the input image include the lengths of various parts of the face included in the input image And the ratio between them.

Description

[0001] FACE RECOGNITION SYSTEM AND METHOD [0002]

The present invention relates to an image recognition method, and more particularly, to a face recognition system and method for finding geometrically similar images.

A common face recognition method searches for the similarity between the components of a face. The components of the face may include the position of the eyes, the position of the mouth, the size of the eyes, the size of the mouth, the histogram of the eyes, and the like. When the size of the eyes is used, it may be difficult to determine the similarity depending on whether the eyes are opened or closed. In the case of using the mouth size, it may be difficult to judge the similarity depending on whether the mouth is opened or closed. When the histogram of the eye is used, it may be difficult to judge the similarity based on the sharpness of the face image, the focus of the eyes, and the like.

It is an object of the present invention to provide a face recognition system and method for detecting a similar face image irrespective of facial expression or face size of an input image.

A face recognition method according to the present invention includes the steps of receiving an input image, extracting facial contour data from the input image, calculating feature elements of the input image using the facial contour data, Comparing the calculated values of each of the candidate IDs with stored values for each of the previously stored feature elements; calculating a similarity value of each of the selected candidate IDs according to the result of the comparing; And outputting at least one similar ID among the candidate IDs according to a magnitude relation between a predetermined threshold value and a predetermined threshold value, wherein the feature elements of the input image include at least one of the lengths of the various parts of the face included in the input image Is determined by the ratio.

The face recognition system according to the present invention includes an input unit for receiving an input image, an extracting unit for extracting facial contour data from the input image, and calculating feature elements of the input image using the facial contour data, Comparing the calculated values of each of the feature elements with stored values for each corresponding feature element stored in advance and calculating a similarity value of each of the candidate IDs selected according to the comparison result, And a recognition unit for outputting at least one similar ID out of the candidate IDs according to the magnitude of the threshold value, wherein the feature elements of the input image are determined by a ratio between the lengths of various parts of the face included in the input image .

According to the embodiment of the present invention, it is possible to detect a similar face image regardless of the facial expression or the face size of the input image by using the feature elements calculated based on the geometric ratio of each part of the face included in the input image System and method.

1 is a block diagram illustrating a face recognition system according to an embodiment of the present invention.
FIG. 2 is a view for explaining the face ratio feature element. FIG.
3 is a view for explaining the eye ratio feature element.
4 is a view for explaining the mouth ratio characteristic element.
5 is a view for explaining the facial feature element.
FIG. 6 is an exemplary diagram illustrating a group of feature data stored in a database.
7 is a diagram for explaining the similarity value.
8 is a flowchart illustrating a face recognition method according to an embodiment of the present invention.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and should provide a further description of the claimed invention. Reference numerals are shown in detail in the preferred embodiments of the present invention, examples of which are shown in the drawings. Wherever possible, the same reference numbers are used in the description and drawings to refer to the same or like parts.

Hereinafter, the face recognition method will be used as an example of the image recognition method for explaining the features and functions of the present invention. However, those skilled in the art will readily appreciate other advantages and capabilities of the present invention in accordance with the teachings herein. Further, the present invention may be implemented or applied through other embodiments. In addition, the detailed description may be modified or changed in accordance with the aspects and applications without departing substantially from the scope, technical ideas and other objects of the present invention.

1 is a block diagram illustrating a face recognition system according to an embodiment of the present invention. Referring to FIG. 1, the face recognition system 100 may include an input unit 110, an extraction unit 120, a recognition unit 130, and a database 140.

The input unit 110 may receive an input image. For example, the input unit 110 may be a device that receives an image image such as a camera, a camcorder, an image sensor, or the like. Also, the input unit 110 may be a user interface for receiving image data. The input unit 110 may transmit the received input image to the extraction unit 120.

The extracting unit 120 can specify a face part in the input image transmitted from the input unit 110. [ The extracting unit 120 may extract the facial contour image from the input image. For example, the extraction unit 120 may extract a facial contour image using a Sobel mask, a Prewitt mask, a Roberts mask, or a Laplacian mask . However, the face contour extraction method is not limited to this. In addition, the extracting unit 120 may remove noise from the extracted face contour image through the filter.

The extracting unit 120 may calculate at least one feature element using the face contour image of the input image. For example, the extracting unit 120 may calculate the first through fourth characteristic elements. The first feature element is the face ratio feature element. The face ratio feature element is a value corresponding to the relationship between the horizontal length and the vertical length of the face contour image. The second feature element is an eye ratio feature element. The eye ratio feature element is a value corresponding to the relationship between the width of the face contour image and the width of the two eyes. The third feature element is the mouth ratio feature element. The mouth ratio feature element is a value corresponding to the relationship between the horizontal length of one eye and the horizontal length of the mouth. The fourth feature element is a face feature element. The facial feature element is a value corresponding to the relationship between the width of the facial contour image and the width of the jaw. The above-mentioned characteristic elements will be described in detail in FIG. 2 to FIG.

The recognition unit 130 may compare the calculated values of at least one feature element of the input image with the stored values of the corresponding feature element in the database 140. [ The database 140 may store a plurality of IDs. The database 140 may store the stored values of the feature elements corresponding to each ID (ID). The recognition unit 130 may compare the calculated value of each feature element of the input image with the stored values of each feature element of the identifiers stored in the database 140. [ For example, the recognition unit 130 may compare the calculated value of the first feature element of the input image with the stored values of the first feature element of the identities (IDs) stored in the database 140. The recognition unit 130 may compare the calculated value of the second feature element of the input image with the stored values of the second feature element of the identities ID stored in the database 140. [ The recognition unit 130 may compare the calculated value of the third feature element of the input image with the stored values of the third feature element of the identities ID stored in the database 140. [ The recognition unit 130 may compare the calculated value of the fourth characteristic element of the input image with the stored values of the fourth characteristic element of the identities stored in the database 140. [

The recognition unit 130 may sequentially compare the calculated values of at least one feature element of the input image with the stored values of the feature elements of the identities (IDs) stored in the database 140. For example, the recognition unit 130 may compare the first feature element for the first time. The recognition unit 130 may select the IDs that include the difference between the calculated value and the stored value within a predetermined similar range according to the comparison result of the first feature element. For example, IDs having a stored value of the first feature element within a predetermined deviation centered on the calculated value of the first feature element of the input image among the IDs may be selected. The IDs selected below are referred to as first candidate IDs. The recognition unit 130 may then compare the calculated values of the second feature elements of the input image with the stored values of the second feature element of the first candidate identities (IDs). As with the first feature element, the recognition unit 130 may include a feature value storage unit having a stored value of the second feature element within a predetermined deviation centered on the calculated value of the second feature element of the input image, among the first candidate IDs 2 Candidate IDs can be selected. The third and fourth candidate IDs (IDs) can be selected by sequentially comparing the third and fourth feature elements in the same manner. Additionally, if there is no selected second candidate ID, the first candidate IDs may be used as the fourth candidate IDs. Also, if there is no selected third candidate ID, the second candidate IDs may be used as the fourth candidate IDs.

The recognition unit 130 may calculate the similarity value of each of the fourth candidate IDs. For example, the recognition unit 130 may calculate the absolute value of the difference between the calculated value of the first feature element of the input image and the stored value of the first feature element of each fourth candidate ID (ID). The recognition unit 130 may calculate an absolute value of the difference between the calculated value of the second characteristic element of the input image and the stored value of the second characteristic element of each fourth candidate ID (ID). The recognition unit 130 may calculate the absolute value of the difference between the calculated value of the third feature element of the input image and the stored value of the third feature element of each fourth candidate ID. The recognition unit 130 may calculate an absolute value of the difference between the calculated value of the fourth characteristic element of the input image and the stored value of the fourth characteristic element of each fourth candidate ID (ID). The recognition unit 130 may multiply the absolute value corresponding to the difference of each feature element by a weight. The similarity value is defined as the sum of the absolute values corresponding to the differences of the first to fourth characteristic elements multiplied by the weights.

The recognition unit 130 may select at least one similar ID (ID) by comparing the similarity value of each of the fourth candidate IDs with a predetermined threshold value. For example, when there is a fourth candidate ID having a similarity value equal to or less than a threshold value, the recognizing unit 130 stores fourth candidate IDs having a similarity value equal to or less than the threshold value, ). If there is no fourth candidate ID having a similarity value equal to or less than the threshold value, the recognizing unit 130 sets the ID having the smallest similarity value among the fourth candidate IDs It can be selected by ID (ID). The recognition unit 130 may output a similar image corresponding to at least one similar ID (ID) and a similar ID (ID) selected.

The database 140 may store various feature data groups. For example, the database 140 may store a feature data group of a smile image image, a feature data group of a multi-image image, a feature data group of an opened image, and the like. In addition, the database 140 may store the feature data group of the past image and the feature data group of the current image. In addition, the database 140 may store the feature data group of the image photographed above the front and the feature data group of the image photographed below the front face, and the like. Each feature data group may include a plurality of IDs. Each feature data group may include at least one feature element corresponding to each ID (ID). The ID (ID) may be superimposed on a plurality of feature data groups. That is, the same ID (ID) may be included in different feature data groups. For example, the same ID (ID) may correspond to an image of the same person. The database 140 may store at least one image corresponding to each ID (ID). In addition, the database 140 may include ancillary data such as a name of a person corresponding to each ID, a resident registration number, and the like.

Initially, the database 140 may be configured by obtaining the first through fourth characteristic elements of the facial image of recognized characters. For example, the calculated values of the first to fourth characteristic elements of the recognized facial image of each person and their corresponding IDs can be stored. Next, a new face image of the recognized person can be input. The calculated values of the first to fourth characteristic elements of the new face image can be compared with the stored values of the database 140. [ As a result of comparison, when the person of the new face image and the person of the similar ID (ID) are the same person, the calculated values of the new face image are excluded and not stored. If the character of the new face image is different from the character of the new face image, the calculated values of the first to fourth characteristic elements of the new face image are additionally stored (stored) in the ID corresponding to the recognized person . In addition, when the data amount of the database 140 becomes equal to or larger than the predetermined capacity, the calculated values of the image that has been excluded can be additionally stored in the database 140. [

The face recognition system 100 according to the present invention can search similar images using the feature elements of the input image calculated based on the geometric ratio between the facial parts. Accordingly, the face recognition system 100 can recognize a similar face regardless of the facial expression. In addition, the face recognition system 100 can recognize a similar face regardless of the face size in the image.

FIG. 2 is a view for explaining the face ratio feature element. FIG. Referring to FIG. 2, the face ratio feature element is a value showing the relationship between the horizontal length and the vertical length of the face. For example, the face horizontal length (A) is defined as the number of pixels between the extension line passing through the center of each pupil and the two points where the face contour meets. However, the outline of the ear is excluded at this time. The face vertical length (B) is defined as the number of pixels between the extension line between two pupils and the center point of the jaw. The vertical length (C) of the mouth is defined as the number of pixels between the straight line passing through the extension line between the two pupils and the center point of the jaw, the outline of the upper lip, and the point where the outline of the lower lip meets. The face ratio feature element is defined as a value obtained by dividing the face vertical length (B) minus the mouth vertical length (C) by the face horizontal length (A). That is, the face ratio feature element is defined as (B-C) / A. The face ratio feature element may be designated as the first feature element.

3 is a view for explaining the eye ratio feature element. Referring to FIG. 3, the eye ratio feature element is a value showing the relationship between the horizontal length of the face and the horizontal length of the two eyes. For example, the face horizontal length (A) is defined as the number of pixels between two points where the extension line passing through the center of each pupil and the face outline meet as shown in Fig. The first eye lateral length (D) is defined as the number of pixels between the extension line passing through the center of each pupil and the two points where the outline of the right eye meets. The second eye lateral length (E) is defined as the number of pixels between the extension line passing through the center of each pupil and the two points where the outline of the left eye meets. The eye ratio feature element is defined as a value obtained by dividing the face lateral length (A) by the sum of the first and second eye lateral lengths (D, E). That is, the eye ratio feature element is defined as A / (D + E). The eye ratio feature element may be designated as a second feature element.

4 is a view for explaining the mouth ratio characteristic element. Referring to Fig. 4, the mouth ratio feature element is a value showing the relationship between the width of the mouth of one eye and the width of the mouth. For example, the first eye lateral length D is defined as the number of pixels between two points where an extension line passing through the center of each pupil and an outline of the right eye meet, as shown in Fig. The mouth width (F) is defined as the number of pixels between the mouths of both eyes. The mouth ratio feature element is defined as the value obtained by dividing the first eye width (D) by the mouth width (F). That is, the input ratio characteristic element is defined as D / F. The mouth ratio feature element may be designated as a third feature element.

5 is a view for explaining the facial feature element. Referring to FIG. 5, the face feature element is a value showing the relationship between the width of the face contour image and the width of the jaw. That is, the facial feature element can show the degree of facial angulation. For example, the face horizontal length (A) is defined as the number of pixels between two points where the extension line passing through the center of each pupil and the face outline meet as shown in Fig. The jaw transverse length (G) is defined as the number of pixels between the extension line passing through both mouths and the point where the face contour meets. The face shape feature element is defined as a value obtained by dividing the face width (A) by the width (G) of the jaw. That is, the facial feature element is defined as A / G. The face shape feature element may be designated as a fourth feature element.

FIG. 6 is an exemplary diagram illustrating a group of feature data stored in a database. Referring to FIG. 6, the database may store a feature data group including first through fourth feature elements. The database 140 shown in FIG. 1 may store various feature data groups. For example, the database 140 may store a feature data group of a smile image image, a feature data group of a multi-image image, a feature data group of an opened image, and the like. Each feature data group may include a plurality of IDs. Each feature data group may include at least one feature element corresponding to each ID (ID). The ID may be included in a plurality of feature data groups redundantly. That is, the same ID (ID) may be included in different feature data groups. For example, the same ID (ID) may correspond to an image of the same person. The database 140 may store at least one image corresponding to each ID (ID).

Each feature data group may include a plurality of IDs. Each of the identifiers id1 to idn may have corresponding first to fourth characteristic elements. For example, the first ID id1 may have storage values a1, b1, c1, d1 of the first through fourth characteristic elements. The second identifier id2 may have storage values a2, b2, c2, d2 of the first through fourth feature elements. The n-th id (idn) may have storage values (an, bn, cn, dn) of the first through fourth characteristic elements. Each feature element may correspond to the feature elements described in Figures 2-5. For example, the first feature element may correspond to the face ratio feature element of FIG. The second feature element may correspond to the eye ratio feature element of Fig. The third feature element may correspond to the fill ratio feature element of Fig. The fourth feature element may correspond to the face feature element of Fig.

7 is a diagram for explaining the similarity value. The recognition unit 130 shown in FIG. 1 may calculate the similarity value of each of the fourth candidate IDs. The recognition unit 130 may output at least one similar ID by comparing the similarity values of the candidate IDs with a predetermined threshold value.

Referring to FIG. 7, the recognition unit 130 receives the calculation values INa, INb, INc, and INd of each feature element of the input image IN and the stored values a2 and b2 of each feature element of the candidate ID id2 , c2, d2) can be calculated. For example, the recognition unit 130 can calculate the absolute value (| INa-a2 |) of the difference between the input image IN and the first feature elements INa and a2 of the candidate ID id2. The recognizing unit 130 may calculate the absolute value of the difference between the input image IN and the second characteristic elements INb and b2 of the candidate ID id2 | INb-b2 |. The recognition unit 130 can calculate the absolute value of the difference between the input image IN and the third characteristic elements INc and c2 of the candidate ID id2 | INc-c2 |. The recognition unit 130 may calculate the absolute value of the difference between the input image IN and the fourth characteristic elements INd and d2 of the candidate ID id2 | INd-d2 |.

In addition, the recognition unit 130 may multiply the absolute value of the difference of each feature element by the weight W. [ The weight W may be applied differently for each feature element. For example, the absolute value of the first feature element difference may be multiplied by the weight w1. The absolute value of the second feature element difference can be multiplied by the weight w2. Third feature The absolute value of the element difference can be multiplied by the weight w3. The absolute value of the fourth feature element difference can be multiplied by the weight w4.

The weight W may be determined by reflecting the degree of dispersion of the first to fourth feature elements in each feature data group of the database 140. [ For example, when the difference between the maximum value and the minimum value of the stored values of the first feature element is 2 (when the difference between the maximum value and the minimum value of a1, a2, ..., an is 2 in FIG. 6) ) May be 0.5. If the difference between the maximum value and the minimum value of the stored values of the second feature element is 1, the weight value w2 may be 1. If the difference between the maximum value and the minimum value of the stored values of the third feature element is 0.5, the weight w3 may be 2. When the difference between the maximum value and the minimum value among the stored values of the fourth feature element is k, the weight w4 may be 1 / k.

The similarity value S is defined as a sum of absolute values corresponding to the difference of each of the first through fourth characteristic elements multiplied by the weight W. [ That is, in the case of the candidate ID id2 in FIG. 7, the similarity value S is expressed by Equation (1).

INa is the calculated value of the first feature element of the input image. INb is the calculated value of the second characteristic element of the input image. INc is the calculated value of the third feature element of the input image. INd is the calculated value of the fourth characteristic element of the input image. and a2 is the stored value of the first feature element of the selected candidate ID id2. and b2 is the stored value of the second characteristic element of the selected candidate ID id2. and c2 is the stored value of the third characteristic element of the selected candidate ID id2. and d2 is the stored value of the fourth characteristic element of the selected candidate ID id2. w1 is a weight (W) corresponding to the first feature element. and w2 is a weight (W) corresponding to the second feature element. and w3 is a weight (W) corresponding to the third feature element. and w4 is a weight (W) corresponding to the fourth characteristic element. The candidate ID having the small similarity value S may be an ID similar to the input image.

8 is a flowchart illustrating a face recognition method according to an embodiment of the present invention. Referring to FIGS. 1 and 8, the face recognition system 100 may output an ID of an image similar in geometry to an input image in a database based on facial feature elements included in an input image.

In step S110, the face recognition system 100 may receive an input image. For example, the face recognition system 100 can shoot an image image in real time. Or the face recognition system 100 may receive image data that has already been photographed.

In step S120, the face recognition system 100 may extract a face contour included in the input image. For example, the face recognition system 100 can extract facial contour images using a Sobel Mask, a Prewitt Mask, a Roberts Mask, or a Laplacian Mask. have. However, the face contour extraction method is not limited to this. In addition, the face recognition system 100 may remove noise from the facial contour image extracted through the filter.

In step S130, the face recognition system 100 can calculate a plurality of feature elements of the input image using the extracted face contours. For example, the face recognition system 100 may calculate the first through fourth feature elements. The first to fourth feature elements may correspond to the face ratio feature element, the eye ratio feature element, the mouth ratio feature element, and the face shape feature element described in Figs. 2 to 5, respectively.

In step S140, the face recognition system 100 may compare the calculated value of each of the feature elements of the input image with the stored values of the corresponding feature element stored in advance. For example, the face recognition system 100 may sequentially compare each feature element. The face recognition system 100 may first compare the calculated values of the first feature elements of the input image with the stored values of the first feature element of the identities ID stored in the database 140. [ The face recognition system 100 is configured to recognize a first candidate having a stored value of a first feature element within a predetermined deviation centered on a calculated value of a first feature element of an input image among IDs stored in the database 140 IDs can be selected. The face recognition system 100 may then compare the calculated values of the second feature elements of the input image with the stored values of the second feature element of the first candidate identifiers (IDs). Similar to the first feature element, the face recognition system 100 may have a stored value of the second feature element within a predetermined deviation centered on the calculated value of the second feature element of the input image among the first candidate IDs The second candidate IDs (IDs) can be selected. The third and fourth candidate IDs (IDs) can be selected by sequentially comparing the third and fourth feature elements in the same manner.

In step S150, the face recognition system 100 may calculate the similarity values of the candidate IDs selected according to the comparison result in step S140. For example, the face recognition system 100 may calculate the similarity value of each of the fourth candidate IDs (IDs). The face recognition system 100 may calculate the absolute value of the difference between the calculated value of the first feature element of the input image and the stored value of the first feature element of each fourth candidate ID. The face recognition system 100 may calculate the absolute value of the difference between the calculated value of the second feature element of the input image and the stored value of the second feature element of each fourth candidate ID (ID). The face recognition system 100 may calculate the absolute value of the difference between the calculated value of the third feature element of the input image and the stored value of the third feature element of each fourth candidate ID (ID). The face recognition system 100 may calculate the absolute value of the difference between the calculated value of the fourth feature element of the input image and the stored value of the fourth feature element of each fourth candidate ID (ID). The face recognition system 100 may multiply the absolute value corresponding to each feature element by a weight. The similarity value is defined as a sum of absolute values corresponding to the first to fourth characteristic elements multiplied by the weight. The similarity value can be calculated by the method described in Fig.

In step S160, the face recognition system 100 may output at least one similar ID according to the magnitude relation between the similarity values of the selected candidate IDs and a predetermined threshold value. For example, if there is a fourth candidate ID (ID) having a similarity value equal to or less than a threshold value, the face recognition system 100 can output IDs having a similarity value equal to or less than the threshold value. If there is no fourth candidate ID having a similarity value equal to or less than the threshold value, the face recognition system 100 sets an ID having the smallest similarity value among the fourth candidate IDs Can be output.

The database 140 may store various feature data groups. For example, the database 140 may store a feature data group of a smile image of the same person, a feature data group of a multi-image image, a feature data group of an opened image, and the like. If the input image includes a smiley face, there is a high probability that the stored values included in the feature data group of the smiley face image of the same person are similar to the calculated values of the input image. In the case where the input image includes a face with a mouth open, there is a high probability that the stored values included in the feature data group of the face image of the same person's mouth are similar to the calculated values of the input image. When the input image includes multiple faces, there is a high probability that the stored values included in the feature data group of the multi-face image of the same person are similar to the calculated values of the input image. However, the face recognition system 100 can finally output the same ID regardless of the feature data group. Further, according to the face recognition method of the present invention, it is possible to output an ID of a person having a geometrically similar ratio even if the person is not the same person. In this case, the ID of the same person is also highly likely to have a degree of similarity below the threshold value.

The face recognition method according to the present invention can search similar images using the feature elements of the input image calculated based on the geometric ratio between the facial parts. Therefore, the face recognition method according to the present invention can search similar faces regardless of the facial expression. In addition, the face recognition method according to the present invention can search similar faces regardless of the face size in the image.

As described above, an optimal embodiment has been disclosed in the drawings and specification. Although specific terms have been employed herein, they are used for purposes of illustration only and are not intended to limit the scope of the invention as defined in the claims or the claims. Therefore, those skilled in the art will appreciate that various modifications and equivalent embodiments are possible without departing from the scope of the present invention. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

100: Face Recognition System
110: input unit
120:
130:
140: Database

Claims (18)

Receiving an input image;
Extracting facial contour data from the input image;
Calculating feature elements of the input image using the face contour data;
Comparing the calculated value of each of the feature elements of the input image with the stored values of the corresponding feature element previously stored;
Calculating a similarity value of each selected candidate ID according to the result of the comparing step; And
And outputting at least one similar ID among the candidate IDs according to a magnitude relation between a similarity value of each of the candidate IDs and a predetermined threshold value,
Wherein the feature elements of the input image are determined by a ratio between the lengths of various parts of the face included in the input image. The method according to claim 1,
Wherein the feature elements of the input image include first through fourth feature elements,
Wherein the first feature element has a first length between two points where an extension line connecting the centers of two eyes and a face contour meet and a second length between the extension line connecting the centers of the two eyes and the center point of the outline of the jaw, Lt; / RTI >
Wherein the second feature element is related to a ratio between a first length and a third length which is a sum of a lateral length of the right eye and a lateral length of the left eye,
Wherein the third feature element is related to a ratio between the transverse length of the right eye and the transverse length of the mouth,
Wherein the fourth feature element is related to a ratio between an extension extending through the first length and both tail edges and a fourth length between two points where the face contour meets. 3. The method of claim 2,
Comparing the calculated values of each of the feature elements of the input image with stored values corresponding to the previously stored corresponding feature element, wherein the calculated value of the first feature element of the input image is compared with the stored first feature element And selects a first candidate ID having a stored value corresponding to the first feature element within a predetermined similar range. The method of claim 3,
Comparing the calculated value of each of the feature elements of the input image with stored values corresponding to the previously stored corresponding feature element to calculate a calculated value of the second feature element of the input image as the second And selecting second candidate IDs having the stored value corresponding to the second feature element within the predetermined similar range from among the first candidate IDs. 5. The method of claim 4,
Comparing the calculated value of each of the feature elements of the input image with stored values corresponding to the previously stored corresponding feature element to calculate a calculated value of the third feature element of the input image as the third And selecting third candidate IDs having the stored value corresponding to the third feature element within the predetermined similar range from among the second candidate IDs. 6. The method of claim 5,
Comparing the calculated value of each of the feature elements of the input image with stored values corresponding to the previously stored corresponding feature element to calculate a calculated value of the fourth feature element of the input image as the fourth And selecting fourth candidate IDs having the stored value corresponding to the fourth feature element within the predetermined similar range from among the third candidate IDs. The method according to claim 6,
Wherein the similarity value is calculated based on the stored values of each of the fourth candidate IDs and the first through fourth characteristic elements of the input image in the step of calculating the similarity values of the selected candidate IDs according to the result of the comparing step. Wherein the calculated face value is calculated based on a difference between the calculated values of the two face images. The method according to claim 1,
Outputting at least one similar ID among the candidate IDs, if there is at least one ID among the candidate IDs whose similarity value is equal to or less than the predetermined threshold value, Way. The method according to claim 1,
And outputting at least one similar ID among the candidate IDs, if there is no ID among the candidate IDs equal to or less than the predetermined threshold value, The face recognition method comprising: The method according to claim 1,
Outputting at least one similar ID among the candidate IDs, and outputting an image corresponding to the at least one similar ID. An input unit for receiving an input image;
An extracting unit for extracting face contour data from the input image; And
Calculating feature elements of the input image using the face contour data, comparing the calculated values of each of the feature elements of the input image with stored values of the corresponding feature elements stored in advance, And a recognition unit for calculating at least one similar ID among the candidate IDs according to a degree of similarity between each of the candidate IDs and a predetermined threshold value,
Wherein the feature elements of the input image are determined by a ratio between the lengths of various parts of the face included in the input image. 12. The method of claim 11,
Wherein the similarity value is calculated based on differences between stored values corresponding to the feature elements of each of the candidate identities and calculated values of the feature elements of the input image. 12. The method of claim 11,
Wherein the recognizing unit outputs the at least one ID when at least one ID among the candidate IDs is equal to or smaller than the predetermined threshold value. 12. The method of claim 11,
Wherein the recognition unit outputs an ID having the minimum similarity value among the candidate IDs when the ID of the candidate IDs is not equal to or less than the predetermined threshold value. 12. The method of claim 11,
Wherein the recognition unit outputs a similar image corresponding to the at least one similar ID. 12. The method of claim 11,
And a database for storing the stored values for each feature element. 17. The method of claim 16,
The database stores feature data groups corresponding to various facial expressions,
Wherein each of the feature data groups includes stored values for each feature element. 17. The method of claim 16,
Wherein the database stores the stored values for each feature element and a corresponding ID using the previously recognized image.

Images