JP2010033141A - Spectacles detection apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a spectacles detection device for highly precisely and inexpensively detecting reflection points from spectacles from a face image, and identifying the presence of spectacles. <P>SOLUTION: This spectacles detection device is provided with: an image pickup part 51 for imaging an object; a face region detection part 52 for detecting a face image from the image captured by the image pickup part 51; a high luminance value detection part 54 for detecting a luminance value higher than a threshold in a face region detected by the face region detection part 52; and a spectacles determination part 55 for determining that the spectacles exist when the luminance value higher than the threshold is detected, and for determining that the spectacles does not exist when the luminance value higher than the threshold is not detected by the high luminance value detection part 54. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a spectacle detection device that detects whether or not a subject wears spectacles based on a reflection point on a human face image.

  An image processing apparatus (see, for example, Patent Document 1) that extracts a reflection point from a spectacle as a feature point on the face in a face image when a person wearing spectacles or the like is illuminated is known. .

  This image processing apparatus includes an image obtained when a person is illuminated (hereinafter referred to as “on image”) and an image obtained when no illumination is illuminated (hereinafter referred to as “off image”). In the difference image, the reflection point from the glasses and the reflection point from the pupil are detected by threshold processing. Then, by performing threshold processing on the ON image, the previously detected reflection point from the glasses and the reflection point from the pupil are separated and identified.

  FIG. 10A shows the on-image, and FIG. 10B shows the brightness value of the on-image. As shown in FIG. 10A, a reflection point 11 from the pupil and a reflection point 12 from the glasses are observed. As shown in FIG. 10B, the luminance value Y1 of the reflection point 11 from the pupil and the luminance value Y2 of the reflection point 12 from the glasses are the luminance value Y0 of the region (background) excluding those reflection points. taller than. In general, the luminance value Y2 of the reflection point 12 from the glasses is higher than the luminance value Y1 of the reflection point 11 from the pupil.

  FIGS. 11A and 12A are diagrams illustrating the difference between the brightness value of the on-image in FIG. 11B and the brightness value of the off-image in FIG. 11C. 11A and 12A, the luminance value Y5 is the luminance value Y1 of the reflection point 11 from the pupil of FIG. 11B and the luminance value Y1 of the reflection point 11 from the pupil of FIG. The brightness value Y6 is the difference between the brightness value Y2 of the reflection point 12 from the glasses in FIG. 11B and the brightness value Y2 of the reflection point 12 from the glasses in FIG. In the difference images shown in FIGS. 11A and 12A, by performing binarization processing using the first threshold Th5 shown in FIG. 12A, the pupil as shown in FIG. The reflection point 11 from the eyeglasses and the reflection point 12 from the eyeglasses can be detected.

  FIG. 13A is a diagram showing the brightness value of the on-image. FIG. 13A shows the brightness value of the reflection point 12 from the glasses that is higher than the brightness value Y1 of the reflection point 11 from the pupil. The reflection point 12 from the glasses can be detected using the binarized image calculated using the second threshold Th4 lower than Y2.

A set of reflection points 12 from the glasses detected using the on-image (FIG. 13) from a set of reflection points 11 from the pupils and the reflection points 12 detected from the glasses previously detected using the difference image (FIG. 11). By removing, the reflection point 11 from the pupil and the reflection point 12 from the glasses can be separated and identified.
JP-A-8-185503

  However, the image processing apparatus disclosed in Patent Document 1 has the following two problems. The first problem will be described. In the difference image (FIG. 11A), the difference luminance value (Y6 in FIG. 11A) due to the reflection point from the glasses may be small depending on disturbance light and illumination conditions. For this reason, it is difficult to detect the reflection point from the glasses by threshold processing of the difference image, and this is the first problem.

  The second problem will be described. In the on-image (FIG. 11B), depending on the ambient light and illumination conditions, the luminance value of the reflection point from the pupil (Y1 in FIG. 11B) and the luminance value of the reflection point from the glasses (FIG. 11B) There are cases where the difference from Y2) is smaller than the luminance value of the reflection point from the pupil. For this reason, it is difficult to separate the reflection point from the pupil and the reflection point from the glasses by performing threshold processing in the on-image, which is a second problem.

  The present invention has been made to solve the above problems, and provides a spectacle detection device that accurately detects a reflection point from spectacles from a face image at low cost and identifies the presence or absence of spectacles. For the purpose.

  In order to achieve the above object, the eyeglass detection device according to claim 1 includes an imaging unit that captures an image of a subject, a face region detection unit that detects a face region from an image captured by the imaging unit, and the face region detection. A high luminance value detecting means for detecting a luminance value higher than a threshold value in the face area detected by the means, and when a luminance value higher than the threshold value is detected by the high luminance value detecting means, it is determined that glasses are present, Glasses determining means for determining that glasses are not present when a luminance value higher than the threshold value is not detected.

  According to the first aspect of the present invention, it is possible to identify the reflection point from the eyeglasses and the reflection point from the pupils by simple threshold processing of only the off-image without reflection from the pupil, and determine the presence or absence of the eyeglasses. At the same time, since the lighting device is unnecessary, the cost can be reduced.

  The eyeglass detection device according to claim 2, wherein the irradiating means for irradiating the subject with light under an irradiation condition in which reflection from the pupil occurs, the imaging means for imaging the subject, and the imaging in a state where the irradiating means irradiates light. A face area detecting means for detecting a face area based on an on-image captured by the means, or an off-image captured by the imaging means in a state where no light is emitted by the irradiating means, the on-image, and the off-image A sum image generating means for generating a sum image composed of the sum of corresponding pixel values with the image, and a high value for detecting a luminance value higher than a threshold in the face area of the sum image generated by the sum image generating means. When a luminance value higher than the threshold is detected by the luminance value detection means and the high luminance value detection means, it is determined that glasses are present, and when a luminance value higher than the threshold is not detected And a, and glasses determining means determines that the mirror is not present.

  According to the second aspect of the present invention, it is possible to identify the reflection point from the glasses and the reflection point from the pupil by simple threshold processing of the sum image of the on image and the off image, and determine the presence or absence of the glasses.

  The eyeglass detection apparatus according to claim 3, wherein the imaging unit is configured to irradiate a subject with light under an irradiation condition in which reflection from a pupil occurs, an imaging unit that images the subject, and the imaging in a state where the irradiation unit irradiates light. A face area detecting means for detecting a face area from an image captured by the means, a high brightness value detecting means for detecting a brightness value higher than a threshold in the face area detected by the face area detecting means, and the high brightness value. When three or more luminance values higher than the threshold are detected by the detection means, it is determined that glasses are present, and when no luminance values higher than the threshold are detected or when two or less luminance values are detected, no glasses are present. Eyeglass determination means for determining

  According to the third aspect of the present invention, it is possible to determine the presence or absence of glasses without identifying the reflection point from the glasses and the reflection point from the pupil.

  As described above, according to the present invention, it is possible to detect a reflection point from spectacles from a face image with high accuracy and low cost, and to easily identify the presence or absence of spectacles.

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

[First Embodiment]
FIG. 1 is a block diagram showing the main configuration of the eyeglass detection device according to the first embodiment of the present invention. As shown in the figure, the eyeglass detection device according to the present embodiment includes an imaging unit 51, a face area detection unit 52, a high luminance value detection unit 54, and a glasses determination unit 55.

  The imaging unit 51 is composed of a CCD camera or the like, and is installed so that the subject who is the subject can be imaged from almost the front, and images an area including the face of the subject.

  The face area detection unit 52 detects the face area of the subject from the image data captured by the imaging unit 51. The method for detecting the face area from the image data may be any known method such as histogram processing, and this method does not matter in the present embodiment.

  The high brightness value detection unit 54 detects a point having a brightness value higher than a certain threshold in the face area obtained by the face area detection unit 52.

  The glasses determination unit 55 determines whether glasses are present in the face area of the subject based on the detection result of the high luminance value detection unit 54.

  Next, the flow of operation of the eyeglass detection device in the present embodiment will be described along the flowchart shown in FIG.

  First, in step 100, the imaging unit 51 images a region including the subject's face. Since it is assumed that only ambient light exists as the illumination condition, an image (off-image) that does not use the illumination device is generated. In this case, as shown in FIG. 3A, the reflection point 12 from the spectacles is observed by disturbance light, whereas the reflection point 11 from the pupil is not observed.

  In step 110, the face area detection unit 52 detects the face area of the subject from the off-image captured by the imaging unit 51.

  In step 120, the high brightness value detection unit 54 performs binarization processing using the threshold value Th1 shown in FIG. 3B from the face region image detected by the face region detection unit 52, and is higher than the threshold value Th1. Detect the luminance value. As the threshold Th1, a value higher than the luminance value Y0 of the background and lower than the luminance value Y2 of the reflection point 12 from the glasses is set. By such threshold processing, the reflection point 12 from the glasses showing the luminance value Y2 higher than the threshold Th1 is detected.

  In step 130, the glasses determination unit 55 determines that glasses are present when one or more luminance values higher than the threshold Th1 are detected by the high luminance value detection unit 54, and a luminance value higher than the threshold Th1 is not detected. Is determined to have no glasses.

  As described above, in the present embodiment, since there is no reflection point 11 from the pupil, threshold processing is facilitated. Therefore, it can be determined easily and accurately whether glasses are present. Further, since no lighting device is used, there is an advantage that the cost can be reduced.

[Second Embodiment]
FIG. 4 is a block diagram showing the main configuration of the eyeglass detection device according to the second embodiment of the present invention. A sum image generation unit 53 is further provided in the configuration of the second embodiment.

  The sum image generation unit 53 generates a sum image constituted by the sum of corresponding pixel values of two images (an off image and an on image described later) captured by the image capturing unit 51.

  Next, the flow of operation of the eyeglass detection device in the present embodiment will be described along the flowchart shown in FIG.

  First, in Step 104, the imaging unit 51 captures an image of a region including the subject's face in a state where the illumination unit 50 irradiates illumination. As an illumination condition, in order to cause not only reflection from the eyeglasses of the subject but also reflection from the pupil, a near-infrared light is illuminated by installing a light source in the vicinity (coaxial) of the optical axis of the camera. Here, as a well-known method, when a light source is installed coaxially with the optical axis of the camera, when near-infrared light is illuminated to the glasses and eyes, not only the reflection from the glasses but also the reflection from the pupil is observed. (E.g., Y. Ebisawa, S. Satoh, Effectiveness of pupil area detection technique using two light sources and image difference method, in: A. Szeto, R. Rangayan (Eds.), Proc. Of the 15th Annual Internet. Conf. Of the IEEE Eng. In Medicine and Biology Society, San Diego, CA, 1993, pp. 1268-1269.). Thereby, an on-image is generated. In this case, as shown in FIG. 10A, the reflection point 12 from the glasses and the reflection point 11 from the pupil are observed by the illumination light. Moreover, in this Embodiment, the imaging part 51 does not perform illumination by the illumination part 50, but it images in the state where only disturbance light exists, and also produces | generates an off image simultaneously.

  In step 110, the face area detection unit 52 detects the face area of the subject from the on-image or off-image captured by the imaging unit 51.

  In step 115, the sum image generation unit 53 generates a sum image of the on image and the off image captured by the image capturing unit 51. FIG. 6A is a diagram showing the brightness value of the generated sum image. The brightness value of each pixel is the brightness value of the on-image in FIG. 6C and the brightness of the off-image in FIG. Consists of the sum of values. In FIG. 6A, the luminance value Y3 is the sum of the luminance value Y1 of the reflection point from the pupil in FIG. 6C and the luminance value Y1 of the reflection point from the pupil in FIG. Y4 is the sum of the luminance value Y2 of the reflection point from the glasses in FIG. 5C and the luminance value Y2 of the reflection point from the glasses in FIG.

  In step 120, the high luminance value detection unit 54 detects the threshold value Th2 shown in FIG. 6A from the face area image detected by the face area detection unit 52 in the sum image generated by the sum image generation unit 53. Is used to perform a binarization process, and a luminance value higher than the threshold Th2 is detected. As the threshold Th2, a value higher than the luminance value Y3 and lower than the luminance value Y4 is set. By such threshold processing, the reflection point 12 from the glasses showing the luminance value Y4 higher than the threshold Th2 is detected.

  In step 130, the glasses determination unit 55 determines that glasses are present when one or more luminance values higher than the threshold Th2 are detected by the high luminance value detection unit 54, and a luminance value higher than the threshold Th2 is not detected. Is determined to have no glasses.

  As described above, in the present embodiment, the reflection point from the glasses is detected based on the sum image of the on image and the off image.

  In the difference image (FIG. 6B), the luminance value of the reflection point from the glasses of the on-image (Y2 in FIG. 6C) due to the influence of disturbance light or the saturation of the high-brightness camera output due to illumination. And the difference (Y6 in FIG. 6B) between the brightness value of the reflection point from the off-image glasses (Y2 in FIG. 6D) may be small. On the other hand, when the sum image (FIG. 6A) is used, the difference between the luminance value of the reflection point from the glasses (Y4 in FIG. 6A) and the luminance value of the background is the difference image (FIG. In general, the difference between the luminance value of the reflection point from the glasses when using B)) (Y6 in FIG. 6B) and the luminance value of the background is larger. From this, it is possible to detect the reflection point from the glasses more easily when the threshold processing is performed on the sum image than when the threshold processing is performed on the difference image.

When only the ON image (FIG. 6C) is used, depending on the ambient light and illumination conditions, the luminance value of the reflection point from the pupil (Y1 in FIG. 6C) and the luminance value of the reflection point from the glasses Since the difference from (Y2 in FIG. 6C) may be smaller than the luminance value of the reflection point from the pupil, the reflection point from the pupil and the reflection point from the glasses may be separated by threshold processing. Have difficulty. On the other hand, when the sum image is used, the difference between the luminance value from the reflection point from the glasses (Y4 in FIG. 6A) and the luminance value from the reflection point from the pupil (Y3 in FIG. 6A) is The luminance value of the reflection point from the glasses when using the on-image (FIG. 6C) (Y2 in FIG. 6C) and the luminance value of the reflection point from the pupil (Y1 in FIG. 6C) Generally greater than the difference. Therefore, it is easier to detect glasses when threshold processing is performed on a sum image than when threshold processing is performed on an on-image, because it is less affected by fluctuation factors such as ambient light and lighting conditions. It becomes possible to do.
[Third Embodiment]
FIG. 7 is a block diagram showing the main configuration of the eyeglass detection device according to the third embodiment of the present invention.

  Next, the flow of operation of the eyeglass detection device in the present embodiment will be described along the flowchart shown in FIG.

  First, in step 106, since the imaging unit 51 captures an area including the face of the subject in a state where the illumination unit 50 irradiates illumination, an on-image can be obtained. Here, the illumination condition is an illumination condition in which not only the reflection from the glasses but also the reflection from the pupil occurs. In this case, as shown in FIG. 10A, the reflection point 12 from the glasses and the reflection point 11 from the pupil are observed by the illumination light.

  In step 110, the face area detection unit 52 detects the face area of the subject from the ON image captured by the imaging unit 51.

  In step 120, the high luminance value detection unit 54 performs binarization processing from the image of the face area detected by the face area detection unit 52 using the first threshold Th3 shown in FIG. A luminance value higher than a threshold value of 1 is detected. As the first threshold Th3, a value that is higher than the luminance value Y0 of the background and lower than the luminance value Y1 of the reflection point 11 from the pupil is set. By such threshold processing, the reflection point 11 from the pupil showing the luminance value Y1 higher than the first threshold Th3 and the reflection point 12 from the glasses showing the luminance value Y2 are detected.

  In step 130, the glasses determination unit 55 determines that glasses are present when the high luminance value detection unit 54 detects three or more luminance values higher than the first threshold Th3, and is higher than the first threshold Th3. When two or less luminance values are detected, it is determined that no glasses are present. This is because there are two reflection points from the pupil existing in one face area, and when the high luminance value detected by the threshold processing described above is three or more, the reflection point from the glasses is detected. Based on what can be determined.

  As described above, in the present embodiment, since it is not necessary to distinguish between the reflection point 11 from the pupil and the reflection point 12 from the glasses, it can be easily determined whether or not the glasses are present.

  In addition, this invention is not limited to embodiment mentioned above, It can apply also to what changed the design within the range described in the claim.

It is a block diagram which shows the main structures of the spectacles detection apparatus which concerns on 1st Embodiment. It is a flowchart which shows the flow of an effect | action of the spectacles detection apparatus which concerns on 1st Embodiment. (A) is an off image, (B) is a figure which shows the luminance value of an off image. is there. It is a block diagram which shows the main structures of the spectacles detection apparatus which concerns on 2nd Embodiment. It is a flowchart which shows the flow of an effect | action of the spectacles detection apparatus which concerns on 2nd Embodiment. (A) is the luminance value of the sum image of the on image and the off image, (B) is the luminance value of the difference image between the on image and the off image, (C) is the luminance value of the on image, and (D) is the luminance value of the off image. It is a figure which shows a value. It is a block diagram which shows the main structures of the spectacles detection apparatus which concerns on 3rd Embodiment. It is a flowchart which shows the flow of an effect | action of the spectacles detection apparatus which concerns on 3rd Embodiment. (A) is a first threshold value of an on-image, and (B) is an image after binarization by the first threshold value. (A) is an ON image, (B) is a figure which shows the luminance value of an ON image. (A) is a luminance value of a difference image between an on image and an off image, (B) is a luminance value of the on image, and (C) is a diagram illustrating a luminance value of the off image. (A) is a first threshold value of a difference image between an on image and an off image, and (B) is a binarized image based on the first threshold value. (A) is the second threshold value of the on-image, and (B) is an image after binarization by the second threshold value.

Explanation of symbols

DESCRIPTION OF SYMBOLS 50 Illumination part 51 Imaging part 52 Face area detection part 53 Sum image generation part 54 High-intensity value detection part 55 Eyeglass determination part

Claims (3)

Imaging means for imaging a subject;
Face area detection means for detecting a face area from an image captured by the imaging means;
High brightness value detection means for detecting a brightness value higher than a threshold in the face area detected by the face area detection means;
Glasses determining means for determining that glasses are present when a luminance value higher than a threshold is detected by the high luminance value detecting means; and for determining that no glasses are present if a luminance value higher than the threshold is not detected;
An eyeglass detection device comprising: Irradiating means for irradiating the subject with light under irradiation conditions in which reflection from the pupil occurs;
Imaging means for imaging the subject;
A face for detecting a face area based on an on-image captured by the imaging unit in a state where light is irradiated by the irradiation unit or an off-image captured by the imaging unit in a state where light is not irradiated by the irradiation unit Area detection means;
Sum image generation means for generating a sum image constituted by a sum of corresponding pixel values of the on image and the off image;
A high luminance value detecting means for detecting a luminance value higher than a threshold in the face area of the sum image generated by the sum image generating means;
Glasses determining means for determining that glasses are present when a brightness value higher than the threshold is detected by the high brightness value detecting means, and determining that glasses are not present when a brightness value higher than the threshold is not detected; ,
An eyeglass detection device comprising: Irradiating means for irradiating the subject with light under irradiation conditions in which reflection from the pupil occurs;
Imaging means for imaging the subject;
A face area detecting means for detecting a face area from an image imaged by the imaging means in a state where light is emitted by the illuminating means;
High brightness value detection means for detecting a brightness value higher than a threshold value in the face area detected by the face area detection means;
When three or more luminance values higher than the threshold are detected by the high luminance value detection means, it is determined that glasses are present, and when a luminance value higher than the threshold is not detected or two or less locations are detected Glasses determining means for determining that glasses are not present;
An eyeglass detection device comprising:

Images