TWI464690B - An iris image definition estimation system using the astigmatism of the corneal reflection of a non-coaxial light source - Google Patents

Description

System for estimating iris image clarity by astigmatism of corneal reflection by non-axial source

The invention relates to a system for estimating the sharpness of an iris image by using a non-axial light source through astigmatism of corneal reflection, in particular to a meridional virtual image and a vector caused by astigmatism when a virtual image is formed by a non-coaxial light source through corneal reflection. Arc virtual image two virtual image imaging points to estimate the iris image clarity and focus adjustment system.

According to the complex structure of the iris on the eyeball, it is a fibrous membrane. On the whole, the iris refers to the colored tissue around the pupil, and the iris has many tiny undulations and strips, which have a unique structure. The dual characteristics of the iris total about 240 degrees of freedom, about 80 degrees of freedom compared to the facial duality feature, and the fingerprint binary feature is about 20 to 40 degrees of freedom. Therefore, to find the same probability of iris coding, the probability of one is 78, so even if it is the iris of the left and right eyes, there will be differences, and the iris is not easy to change with age. So far, the iris recognition rate is the highest of all biometrics and more difficult to forge than other biometrics.

The area of the human iris is very small, and the diameter is only about one centimeter. Therefore, in order to obtain a clear iris texture, a long focal length sensing element lens must be used, but the long focal length lens has a very shallow depth of field, and must be properly focused during the image capturing process. In order to obtain a clear iris texture, you must use auto focus or guide focus In order to obtain an image that can be recognized by the iris. Whether it is autofocus or guided focus, you need to be able to evaluate two factors: (a) the degree of focus of the image, and (b) the direction of focus adjustment. The conventional image in-focus resolution is roughly the intensity of the high-frequency signal of the measured image. Due to the large amount of calculation, it is disadvantageous for immediate resolution calculation. Although Park & Kim uses a coaxial light source to save computation time, the existing iris focus evaluation technology cannot adjust the focal length from the single image. Therefore, it is necessary to waste time to guess the direction of focus adjustment. Blind guessing and the fact that the user's head can't be kept will cause the auto focus or the guide focus to produce incorrect results, which greatly increases the time required for image acquisition. The current solution is to use the distance measuring device to detect the target head distance to focus on, as in US Pat. No. 7,095,901 B1, but this leads to an increase in the cost of the device, and the head position and the depth of the eye socket and the iris position are caused by people. However, it is not possible to solve the problem of iris focusing straight.

A system for automatically estimating the sharpness of iris images is known as US 5,291,560A (hereinafter referred to as citation 1), which is a widely used iris recognition technology, using a semi-lens combined with a camera and a small screen for use. You can see the iris image taken by the camera. The user adjusts the position of the head to make the iris texture clear. On the one hand, image processing technology is also used to automatically determine whether the image is clear enough. There is no guidance for this process, users need to be trained and used to use it. Images that are judged to be clear will undergo iris localization, normalization, and feature extraction and comparison to identify their identity.

A system for estimating the sharpness of an iris image using image high-frequency signal intensity is, for example, US2004/0101170A1 (hereinafter referred to as citation 2). This citation 2 proposes a method for evaluating whether iris images are clear. First, they detect whether there are upper and lower eyelids in the image. A strong horizontal edge, if the horizontal edge is too weak, the image is determined to be blurry. The initially screened image is then subjected to a simple pupil positioning to remove a reduced image of the iris region image on both sides of the pupil. Both sides of the pupil are less susceptible to eyelids and eyelashes, so using the image gradient intensity of this interval, you can estimate whether the iris texture is clear. However, the problem is that the gradient strength of the iris texture is easy to use, and it is difficult to determine the criteria that are universally applicable. This method is used to determine whether the iris texture is clear.

A system for estimating the sharpness of an iris image using image high-frequency signal intensity is US 6,714,665 B1 (hereinafter referred to as citation 3), which uses a pair of wide-angle cameras to position the user's head and eyes with stereoscopic vision technology. Manipulating a plane mirror orientation allows the user's iris texture of one eye to be projected onto a long focal length camera. In the process of controlling the mirror to search for the eye, the corneal reflections of the two sources are used to confirm whether the eye has been found. This system must be calibrated in advance to use the three cameras to position the eye and use the calculated depth information to approximate the focal length to the eye. In the process of searching for clear focus, the trial and error method is used: first try to adjust the focal length in a certain direction. If the adjusted image becomes more unclear, then the reverse adjustment, so that the adjustment is repeated, the clear image can be obtained. . This system is extremely costly and difficult to calibrate. The movement of the user's head during autofocusing will make focusing time consuming and error prone.

A system for estimating the sharpness of an iris image using image high-frequency signal intensity is, for example, US 6,753,919 B1 (hereinafter referred to as citation 4), which aims to turn the iris recognition system into a hand-held device. In order to allow the user to focus on themselves, they placed a concave mirror that reflects visible light but allows infrared light to pass in front of the camera. The user can see the magnified virtual image of his eye by the visible light reflected by the concave mirror, and the camera can obtain the infrared image of the iris texture and calculate the sharpness of the image through a filter. This device can also be externally attached to a lens, making it easy for users to see their own eye images. The problem is that this device also has no focus guidance system, and the user must blindly guess to zoom in or out of the image capture device to obtain a clear image.

A system for estimating the sharpness of an iris image using a ranging system is, for example, US 7,095,901 B2 (hereinafter referred to as citation 5), which mainly adds an additional light-receiving element to the iris imaging system to accurately position the user's forehead. Position with cheeks to focus calculations. This patent proposes two possible implementations, one of which is to use an additional camera for ranging, and the other is to switch the iris camera in both the ranging and iris imaging modes. The biggest problem with this method is that the distance between the head position and the iris texture is inconsistent, so it can only be regarded as a method of coarse focus, and it is not guaranteed to obtain a clear image.

A system for estimating the sharpness of iris images using a coaxial light source, such as Park and Kim (Kang Ryoung Park and Jaihie Kim, "A Real-Time Focusing Algorithm for Iris Recognition Camera," IEEE Transactions on Systems, Man, and Cybernetics, Vol. 35 , No.3, Pp.441-444, Aug.2005, (hereinafter referred to as citation 6), the citation 6 mainly uses a coaxial light source as an incident light source. After the coaxial incident light is reflected by the cornea, a virtual image is formed behind the cornea and near the iris. When the focal length of the camera is not correct, the virtual image will be symmetrically scattered into a nearly circular bright area due to out-of-focus. Calculate the area of this bright area to know if the iris is clear. However, the technique of Citation 6 can only judge the image clarity and cannot provide the focus adjustment direction, so it is also necessary to use the trial and error method for autofocus. The movement of the user's head during autofocusing will make focusing time consuming and error prone.

It can be seen that there are still many shortcomings in the above-mentioned conventional technology, which is not a good designer, but needs to be improved.

In view of the various shortcomings derived from the above-mentioned various systems for estimating the clarity of iris images, the inventors of the present invention have improved and innovated, and after years of painstaking research, finally succeeded in research and development, the use of non-axial light sources through corneal reflection. A system for estimating the sharpness of iris images by astigmatism.

The object of the present invention is to provide a system for estimating the sharpness of an iris image by using an astigmatic phenomenon of a non-axial source through corneal reflection, and using a non-coaxial light source to directly estimate the image sharpness and the focus adjustment direction from a single image. . It simplifies the calculation required to obtain a clear iris image, achieves the speed of immediate response, and can be applied to iris autofocus or guide the user to move the focus back and forth to obtain a clear iris image for iris recognition.

A second object of the present invention is to provide a corner angle using a non-axial source Membrane reflection astigmatism A system for estimating the sharpness of an iris image. A non-coaxial light source is used to generate an incident light that is incident into the cornea from an oblique angle from the optical axis of the image sensing element for the measured incident light to pass. The amount of out-of-focus displacement after the cornea reflects astigmatism makes the spots at different distances exhibit different bright-area shapes. The biggest feature is the ability to take advantage of changes in the shape of the bright area while assessing the sharpness of the iris image and the direction of focus adjustment.

A system for estimating the sharpness of an iris image by a astigmatic phenomenon of a non-axial source through corneal reflection, comprising: a non-coaxial light source for generating an incident light, the oblique angle of the incident light from the optical axis Injecting into the cornea for measuring the amount of change in the out-of-focus displacement of the incident light after the astigmatism of the cornea, forming a meridional virtual image and a sagittal arc virtual image behind the cornea; for example, if the light source is placed directly below the camera, then The meridional plane bright area of the optical system will be approximately vertical, while the sagittal plane bright area will be approximately horizontal.

The sensing component of the image is obtained by using the sensing component to capture a composite bright region shape of a meridian virtual image and a arc virtual image projected onto the image plane, thereby providing adjustment of the focal length direction, when the focal length of the sensing component is adjusted to the meridional virtual image When it is nearby, it is better to focus in the direction of the meridional plane bright area, and the direction of the bright area of the sagittal plane is more blurred, the focal length of the sensing element is adjusted farther; when the focal length of the sensing element is adjusted near the sagittal arc virtual image, The focus of the bright area of the arc plane is better, and the blur of the bright area of the meridional plane is larger, the focal length of the sensing element is adjusted; and the single sheet of the non-coaxial light source is reflected by the cornea by the sensing element The area occupied by the bright areas in the iris image can be used to estimate the sharpness of the iris image, while the shape of the bright area provides the direction of focus adjustment.

Referring to FIG. 1 , the system for estimating the sharpness of an iris image by using a non-axial source through astigmatism of a corneal reflection includes a non-coaxial light source 1 for generating an incident light 5 , the incident light. 5 is injected into the cornea 9 from an oblique angle from the optical axis 4 for measuring the amount of change in the out-of-focus displacement of the incident light 5 after the astigmatism of the cornea 9 is formed at one of the meridional virtual image 6 and the one sagittal virtual image 7 behind the cornea. Glint shape (outline shape); an image sensing element 2, using the sensing element 2 to capture a Meridional virtual image 6 and a Sagittal virtual image 7 to provide a focal length The adjustment of the direction, when the focal length of the sensing element 2 is adjusted near the meridional virtual image 6, the focusing in the meridional direction is better, and the degree of blurring in the sagittal direction is greater; when the focal length of the sensing element 2 is adjusted near the sagittal virtual image 7, The focus in the sagittal direction is better, and the degree of blur in the meridional direction is larger. When the focus position of the sensing element 2 is different, the meridional virtual image 6 and the sagittal virtual image 7 change the shape of the virtual image 8 of the light source due to different degrees of defocus.

Wherein, the virtual image 8 of the light source occupies an image area size, which can be used to estimate the image focus quality, and the area of the meridian virtual image 6 and the sagittal virtual image 7 can be used to indicate the direction of adjusting the focal length.

Please refer to FIG. 1 , mainly using a non-coaxial light source 1 , and the number of non-coaxial light sources is not limited. When the non-coaxial light source 1 generates an incident light 5 , the incident light 5 From the oblique angle from the optical axis, the cornea 9 is injected, and the incident light 5 passes through the astigmatism displacement of the cornea 9 to make the imaging unclear, and the astigmatism phenomenon (Astigmatism) is formed, and is not formed behind the cornea. Coincident one of the meridian virtual image 6 and a vector arc virtual image 7. The astigmatism causes the original object to become two separate and perpendicular short lines after imaging, and the projections are combined on the image plane to form an elliptical spot. In principle, in short, the meridian virtual image points and the sagittal arc virtual image points do not coincide, resulting in different imaging of different axial positions.

When the distance between the non-coaxial light source 1 and the eyeball is much larger than the radius of curvature of the cornea, even if the head moves slightly, the changes of the two virtual images with respect to the position of the eyeball can be ignored.

The estimation method of focusing degree and focusing direction is summarized as follows: when the focal length of the sensing element 2 is from the side of the meridian virtual image 6 (see Fig. 1, the object distance is small) to the side of the sagittal arc virtual image 7 (see Fig. 1, the object distance) When the adjustment is larger, the captured image sequence is as shown in Figure 2 (in this example, two non-coaxial light sources placed under the sensing element are used, so the meridional plane bright area is about the vertical plane, and the sagittal arc The bright area of the plane is about the horizontal plane.) Because of the two light sources, there are two spots that reflect).

Please refer to Figure 3, because the focal length is adjusted near the meridian virtual image 6, so the meridian (vertical) direction is better focused, and the sagittal arc (horizontal) direction is more blurred; please refer to Figure 4, because the focal length is adjusted in the sagittal virtual image 7 Nearby, so the focus of the sagittal arc is better, while the meridional direction is more blurred. The area occupied by the bright areas in a single image can be used to estimate the sharpness of the iris 3 image, while the bright area The shape provides the direction of focus adjustment, and the sensing element 2 captures the above two information for automatic focusing or guiding focusing.

The virtual image formed by the corneal reflection of the non-coaxial light source 1 is converted into a bright region in the image with high brightness and easy detection by the sensing element 2, and this bright region is deformed by the astigmatism phenomenon. The sensing element 2 focuses on different object distances, and the shape of the bright area will also be different. Since the position of the iris 3 is just near the virtual image 8 position of the light source in the cornea, the change of the shape of the bright area can be used to estimate the sharpness of the texture of the iris 3. If the texture of the iris 3 is not clear enough, the shape of the bright area can be further used to indicate that the image can be obtained. The focus of the clearer image is adjusted. The infrared illumination source and the sensing element 2, which are necessary for directly taking the image of the iris 3, can quickly estimate the texture definition of the iris 3 and the direction of the focal length adjustment without an additional device.

Please refer to FIG. 5 , which is a flow chart of the system for estimating the sharpness of the iris image by the astigmatism phenomenon of the non-axial light source through the cornea reflection according to the present invention: 1. Using the sensing element 2 to extract the incident light emitted by the non-coaxial light source 1 5, for generating an incident light, the incident light is incident into the cornea from an oblique angle from the optical axis, for measuring the amount of defocusing displacement of the incident light after passing through the corneal astigmatism, and forming a meridian virtual image behind the cornea 6 and a sagittal arc virtual image 7, these two virtual images will be projected to the same position of the image sensing element to form a bright area that is easy to detect; second, the light source is extracted through the cornea to reflect the bright area; 3. Estimate the single light source to reflect the bright area outer frame rectangle along the meridian The plane measures the average height (denoted as H) and the measured average width along the sagittal plane (denoted as W). Value; 4. If the rectangle of the light-reflecting bright area is captured, the average height of the meridian plane bright area (H) and the average width of the sagittal plane bright area (W) are within the preset range, then the auto-focus is determined. 5. When the average height (H) of the meridian plane bright area is too large, the focal length of the sensing element 2 is adjusted closer, or the user column is guided back; if the average width (W) of the sagittal plane bright area is too large, then The focal length of the sensing element 2 is adjusted to be farther, or the user is guided to advance; finally, the first step is repeated until the fourth step, the auto focus is completed, that is, the sharpness of the iris 3 image is adjusted.

The system for estimating the sharpness of the iris image by the astigmatism phenomenon of the non-axial light source through the cornea reflection provided by the present invention has the following advantages when compared with the aforementioned cited cases and other conventional techniques:

First, the calculation method is simpler and faster than other conventional techniques. The user does not need to learn how to focus on the image through additional training, which can greatly reduce the time of focus adjustment and the wrong result of focus guidance.

Second, the iris 3 is directly used to identify the sensing element 2 and the light source required by the image capturing system, and the focus guiding system or device is not additionally added, which reduces the burden on the equipment cost.

The detailed description of the preferred embodiments of the present invention is intended to be limited to the scope of the invention, and is not intended to limit the scope of the invention. The case In the scope of patents.

In summary, this case is not only innovative in terms of space type, but also can enhance the above-mentioned multiple functions compared with the customary items. It should fully meet the statutory invention patent requirements of novelty and progressiveness, and apply for it according to law. This invention patent application, in order to invent invention, to the sense of virtue.

1‧‧‧non-coaxial light source

2‧‧‧Sensor components

3‧‧‧Iris

4‧‧‧ optical axis

5‧‧‧Incoming light

6‧‧‧ Meridian virtual image

7‧‧‧Yan arc virtual image

8‧‧‧Light source virtual image

9‧‧‧Cornea

Figure 1 is a schematic diagram of a system for estimating the sharpness of an iris image by a astigmatic phenomenon of corneal reflection by a non-axial source; Figure 2 is a system for estimating the sharpness of an iris image by a astigmatism of a non-axial source through corneal reflection. The bright region change map of the iris image sequence obtained by the non-axial light source after being reflected by the cornea at different focal setting; FIG. 3 is the system for estimating the iris image clarity by the astigmatism phenomenon of the non-axial light source through the cornea reflection The meridional direction is focused on the better image; FIG. 4 is a better view of the sagittal arc focusing of the system for estimating the sharpness of the iris image by the astigmatism of the non-axial source through the cornea; FIG. 5 is the non-axial source through the cornea. Systematic flow chart for estimating the clarity of iris images by the astigmatism of reflection.

1‧‧‧non-coaxial light source

2‧‧‧Sensor components

3‧‧‧Iris

4‧‧‧ optical axis

5‧‧‧Incoming light

6‧‧‧ Meridian virtual image

7‧‧‧Yan arc virtual image

8‧‧‧Light source virtual image

9‧‧‧Cornea

Claims (8)

A system for estimating the sharpness of an iris image by using an astigmatic phenomenon of a non-axial light source through corneal reflection is a cornea incident on a cornea with a light source, and a projection point of the light source is measured by the sensing component as a basis for focusing the light source, including : a non-coaxial light source for generating an incident light that is incident into the cornea from an oblique angle from the optical axis for measuring a change in the out-of-focus displacement of the incident light after passing through the corneal astigmatism, and forming a rear surface behind the cornea The meridian virtual image and the one-vector virtual virtual image; a sensing component that uses the sensing element to capture a composite bright region shape of a meridian virtual image and a vector arc virtual image, thereby providing adjustment of the focal length direction. The scope of the patent application of paragraph 1 of the light source by using a non-axial astigmatism of the cornea reflection image sharpness estimated iris system, wherein the composite image area size occupied by the shape of the bright region, the image can be used to estimate focus good Bad, and the size of the meridian virtual image and the sagittal arc virtual image can be used to indicate the direction of adjusting the focal length. The scope of the patent application of paragraph 1 of the light source by using a non-axial astigmatism of the cornea reflection image sharpness estimated iris system, wherein the meridional shape of the bright region due to the virtual image formed is out of focus bright region meridian plane, the composite light The shape of the area occupies the size of the image area, and the direction of the bright area of the meridional plane is blurred, and the focal length of the sensing element is adjusted. The scope of the patent application of paragraph 1 of the light source by using a non-axial astigmatism of the cornea reflection image sharpness estimated iris system, wherein the vector arc shape of the bright region due to the virtual image formed is out of focus bright area sagittal plane arc, The size of the composite bright area occupies the image area. If the direction of the bright area of the sagittal plane is blurred, the focal length of the sensing element is adjusted far. The scope of the patent application of paragraph 1 of the light source by using a non-axial astigmatism of the cornea reflection image sharpness estimated iris system, wherein the sensing element is not the same in-focus position, because of the different virtual meridional and sagittal arcs out as the virtual image The degree of focus changes the shape of the bright area of the virtual light source of the composite light source. The system for estimating the sharpness of an iris image by the astigmatism phenomenon of the non-axial source through the cornea reflection according to the fifth aspect of the patent application, wherein the image area of the virtual image of the light source can be used to estimate the image focus. The size of the meridian virtual image and the sagittal virtual image can be used to indicate the direction of adjusting the focal length. A system for estimating the sharpness of an iris image by a astigmatism phenomenon of a non-axial source through corneal reflection according to the fifth aspect of the patent application, wherein the shape of the bright region formed by the out-of-focus virtual image is a meridional plane bright region, and the light source virtual image The size of the image area, the degree of blurring of the meridional plane bright area is greater, and the focal length of the sensing element is adjusted closer. A system for estimating the sharpness of an iris image by a astigmatism phenomenon of a non-axial source through corneal reflection according to the fifth aspect of the patent application, wherein the shape of the bright region formed by the out-of-focus image is a sagittal plane bright region. The image area of the virtual image of the light source is large. If the direction of the bright area of the sagittal plane is blurred, the focal length of the sensing element is adjusted far.