JP2007011989A - Image processing apparatus, image processing method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus capable of accurately extracting an object to be authenticated, an image processing method and a program. <P>SOLUTION: A discriminating object out of images obtained as the image pickup result of the discriminating object included in a living body is thickened and then the thickened discriminating object is thinned. Thereby even when a disconnected portion is caused in the discriminating object due to the reduction of a contrast difference between a net-like discriminating object and the other, since the disconnected portion is connected and then the discriminating object is thinned, the net-like discriminating object can be accurately extracted. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an image processing apparatus, an image processing method, and a program, and is suitable for application to biometric authentication.

  Conventionally, many authentication apparatuses that use biometric fingerprints as targets for biometric authentication have been proposed, but in recent years, blood vessels in the living body have attracted attention as one of targets for biometric authentication.

Specifically, an apparatus that extracts a blood vessel by performing edge extraction processing on image data obtained as a blood vessel imaging result and generates a line passing through the center of the blood vessel as identification target data. Has been proposed (see, for example, Patent Document 1).
Japanese Patent Laid-Open No. 10-079034

  By the way, in the apparatus having such a configuration, when the identification target is, for example, a blood vessel in the fundus, it is obtained as an imaging result of the blood vessel because there are relatively few other tissues interposed between the blood vessel and the living body surface. Since the contrast between the blood vessel and the other contrast in the image data to be obtained becomes large, the blood vessel to be noted can be accurately extracted.

  However, when the identification target is, for example, a blood vessel in a finger of a living body, an image obtained as an imaging result of the blood vessel due to a relatively large amount of other tissues interposed between the blood vessel and the surface of the living body. The blood vessel and other contrast differences in the data become smaller. In addition, even when the device that captures the authentication device is inferior, the difference between the identification target and the other contrast in the image data obtained as the imaging result may be small. In such a case, there is a problem that a blood vessel to be focused on cannot be accurately extracted due to a small contrast difference.

  The present invention has been made in consideration of the above points, and an object of the present invention is to propose an image processing apparatus, an image processing method, and a program capable of accurately extracting an identification target.

  In order to solve such a problem, the present invention is an image processing apparatus, comprising: a thickening unit that thickens the identification target among images obtained as a result of imaging a net-like identification target in a living body; and the thickening unit. And a detecting means for detecting the middle line of the identification target that has been bolded.

  Therefore, in this image processing apparatus, even when a gap portion occurs in the identification target due to a small contrast difference between the mesh-like identification target and the other, the gap portion is combined. Therefore, a net-like identification target can be accurately extracted.

  The present invention also relates to an image processing apparatus, an edge detection unit that detects an edge of an image obtained as a result of imaging a net-like identification target in a living body, and a smoothing that smoothes an image detected by the edge detection unit. Of the image smoothed by the smoothing means, the thickening means for thickening the identification target, and the thinning means for thinning the identification target thickened by the thickening means. I made it.

  Therefore, in this image processing apparatus, since smoothing is performed after edge detection, even when an image having a small contrast difference between the mesh-like identification target and the other image is obtained, discontinuity occurs in the image after edge detection. Not only can the dots not be removed to accurately remove the noise that does not belong to the mesh-like identification target, but also because the mesh-like identification target is thickened and then thinned, the contrast difference is reduced. Even when a discontinuous portion occurs in the identification target, the discontinuous portion can be combined and then thinned to accurately extract a net-like identification target.

  Furthermore, the present invention is an image processing method, wherein an image obtained as an imaging result of a net-like identification target in a living body is a first step for thickening the identification target, and the thickened identification target is thinned. A second step is provided.

  Therefore, in this image processing method, even when a gap portion is generated in the identification target due to a small contrast difference between the mesh-like identification target and the other, the gap portion is combined. Since the line is thinned, a net-like identification target can be accurately extracted.

  Furthermore, the present invention is a program, which is a first process for thickening an identification target among the images for a computer responsible for controlling an apparatus that processes an image obtained as an imaging result of the identification target in a living body. And a second process for thinning the thickened identification target.

  Therefore, in this program, even if a gap portion occurs in the identification target due to a small contrast difference between the mesh-like identification target and the other, the thinning is performed after combining the gap portions. Therefore, it is possible to accurately extract the net-like identification target.

  According to the present invention, among the images obtained as an imaging result of the identification target in the living body, the identification target is thickened, and the thickened identification target is thinned. Even if there is a discontinuity in the identification target due to a small contrast difference with the other, the mesh-like identification target is accurately identified because the discontinuity is combined and thinned. Thus, the identification target can be extracted with high accuracy.

  Further, according to the present invention, an edge of an image obtained as a result of imaging a net-like identification target in a living body is detected, the edge-detected image is smoothed, and the identification target is bolded out of the smoothed image, By thinning the bold identification object, even if an image with a small contrast difference between the mesh-like identification object and the others is obtained, there is no problem that occurs in the image after the edge detection. Not only is it possible to accurately remove noise that does not belong to the mesh-like identification target by removing continuous points, but even if there is a discontinuity in the identification target due to the small contrast difference Then, the discontinuous portions are combined and then thinned to accurately extract a net-like identification target, and thus the identification target can be extracted.

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

(1) Configuration of Biological Information Generating Device In FIG. 1, reference numeral 1 denotes the biometric information generating device 1 according to the present embodiment as a whole. An information generation unit 3 that generates data for identifying a living body (hereinafter referred to as biological identification data) is connected to each other via a cable.

(2) Configuration of Blood Vessel Imaging Unit This blood vessel imaging unit 2 has a curved guide groove 11 formed so as to imitate a finger FG at a predetermined position of the housing 1A of the authentication device 1, and the guide groove 11 An imaging opening 12 is disposed on the bottom surface of.

  A colorless and transparent opening cover portion 13 made of a predetermined material is provided on the surface of the imaging opening portion 12, while a camera portion 14 is disposed immediately below the imaging opening portion 12 inside the housing 1A. .

  Further, on the side surface of the guide groove 11, a pair of near-infrared light sources 15 (15 </ b> A and 15 </ b> B) that irradiate near-infrared light specifically absorbed with respect to hemoglobin as imaging light for blood vessels are provided on the guide groove 11. Is arranged so as to sandwich the imaging opening 12 in parallel with the short direction of the finger, and can be irradiated with near-infrared light on the finger pad side portion of the finger FG in contact with the guide groove 11. .

  Therefore, in this blood vessel imaging unit 2, the ratio of near-infrared light reflected on the surface of the finger FG is remarkably suppressed as compared with the case where near-infrared light is irradiated on the finger pad base of the finger FG. Near-infrared light incident on the inside of the finger FG via the finger FG is absorbed by hemoglobin passing through the blood vessel and scattered in tissues other than the blood vessel through the inner side of the finger FG and from the finger FG. As near-infrared light for projecting blood vessels (hereinafter referred to as blood vessel projection light), the light enters the camera unit 14 via the imaging opening 12 and the opening cover 13 in sequence.

  In the camera unit 14, the macro lens 16 and near-red light that transmits only near-infrared light in a wavelength range (approximately 900 [nm] to 1000 [nm]) that depends on both oxygenation and deoxygenation hemoglobin. An external light transmission filter 17 and a CCD image sensor 18 are sequentially arranged, and blood vessel projection light incident from the opening cover 13 is sequentially passed through the macro lens 16 and the near-infrared light transmission filter 17 to the CCD image sensor 18. To the imaging surface. Thus, the camera unit 14 can faithfully image both venous and arterial capillaries mixed inside the adjacent finger FG.

  Under the control of the information generation unit 3, the CCD imaging device 18 images a blood vessel or the like that is imaged on the imaging surface, and outputs the imaging result to the information generation unit 3 as an image signal.

  In this way, the blood vessel imaging unit 2 is capable of imaging blood vessels existing in the finger of the living body.

(3) Configuration of Information Generation Unit On the other hand, the information generation unit 3 is configured by connecting a blood vessel imaging drive unit 21 and an image processing unit 22 to the control unit 20 as shown in FIG. ing.

  The control unit 20 includes a central processing unit (CPU) that controls the entire biological information generating apparatus 1, a ROM (Read Only Memory) that stores various programs, and a RAM (Random Access Memory) as a work memory of the CPU. ), And various instructions are given to the control unit 20.

  When the blood vessel extraction command COM for extracting a blood vessel is given from the operation unit (not shown), the control unit 20 sets the operation mode to the blood vessel extraction mode based on the corresponding program stored in the ROM. Then, the blood vessel imaging driving unit 21 and the image processing unit 22 are controlled.

  In this case, the blood vessel imaging drive unit 21 activates the blood vessel imaging unit 2 so as to drive the near-infrared light source 15 and the CCD imaging device 18 of the camera unit 14 respectively. As a result, in the blood vessel imaging unit 2, near infrared light is irradiated from the near infrared light source 15 onto the finger pad side of the imager's finger FG (FIG. 1) that is in contact with the guide groove 11 (FIG. 1) at this time. Then, the blood vessel projection light guided to the imaging surface of the CCD image sensor 18 via the finger FG (FIG. 1) is output from the CCD image sensor 18 to the image processing unit 22 as a blood vessel projection image signal S1.

  As shown in FIG. 3, the image processing unit 22 includes an A / D (Analog / Digital) conversion unit 31, an edge extraction unit 32, a smoothing processing unit 33, a binarizing unit 34, a thickening unit 35, and a center line. The blood vessel projection image signal S <b> 1 configured by the detection unit 36 and supplied from the CCD image sensor 18 is input to the A / D conversion unit 31.

  The A / D conversion unit 31 performs A / D conversion processing on the blood vessel projection image signal S1, and sends the blood vessel projection image data D1 obtained as a result to the edge extraction unit 32.

  The edge extraction unit 32 sequentially performs, for example, a filter process called Laplacian on the blood vessel projection image data D1. As a result, the contour in the image of the blood vessel projection image data D1 is emphasized more than other elements, and the blood vessel as the contour is extracted. Then, the edge extraction unit 32 sends an image from which blood vessels have been extracted as the filter processing result as data (hereinafter referred to as blood vessel extraction image data) D2 to the smoothing processing unit 33.

  The smoothing processing unit 33 performs a predetermined smoothing process on the blood vessel extraction image data D2. Specifically, as shown in FIG. 4, the smoothing processing unit 33 detects the luminance values of the pixels in the setting range AR centered on the target pixel NP, and detects all the pixels in the setting range AR. An average value (hereinafter referred to as a first average value) is obtained (FIG. 4A), and an average value of each pixel (hereinafter, referred to as “first pixel average value”) excluding the target pixel NP from all pixels in the setting range AR. This is called a second average value) (FIG. 4B).

  Then, when the difference between the first average value and the second average value is less than the predetermined threshold value, the smoothing processing unit 33 determines that the target pixel NP is not noise and the luminance of the target pixel NP. Keep the value as is.

  On the other hand, when the difference between the first average value and the second average value is equal to or greater than a predetermined threshold value, the smoothing processing unit 33 determines that the target pixel NP is noise, and The brightness value of NP is replaced with the first average value or the second average value.

  Even when there is a noise component pixel around the pixel of interest NP, the pixel of interest NP is not erroneously detected as noise due to the influence of the pixel. This is because, even if a pixel other than the target pixel NP is a noise component among all the pixels in the setting range AR, the first average value and the second average are determined if the target pixel NP itself is not noise. This is because the difference in values never exceeds a predetermined threshold.

  In this way, the smoothing processing unit 33 performs the above-described smoothing processing in a predetermined order with a predetermined pixel in the image IM as the target pixel NP, and converts the smoothed image into data (hereinafter referred to as smoothing). Sent to the binarization section 34 as D3).

  The binarization unit 34 performs binarization processing on the smoothed blood vessel extraction image data D3. As a result, in the image of the blood vessel extraction image data D2, a pixel having a luminance value equal to or lower than the predetermined luminance value is converted as a pixel having the lowest luminance value (hereinafter referred to as a white pixel), and pixels larger than the predetermined luminance value are displayed. This is converted as a pixel having the highest luminance value (hereinafter referred to as a black pixel), and only the blood vessel is left in the image. Then, the binarization unit 34 sends an image in which only the blood vessel is left as a result of the binarization processing to the thick line unit 35 as data (hereinafter referred to as blood vessel image data) D4.

  The thick line unit 35 performs, for example, a filtering process called dilution on the blood vessel image data D4. As a result, the blood vessels in the image of the blood vessel image data D4 are thickened, and the portions that were originally disconnected but connected are connected. Then, the thickening unit 35 sends an image in which the blood vessel is thickened as the filter processing result to the centerline detection unit 36 as data (hereinafter referred to as thick blood vessel image data) D5.

  The center line detection unit 36 performs a filtering process called Thinning, for example, on the large blood vessel image data D5. As a result, the blood vessel in the image of the thick blood vessel image data D5 is thinned, and the blood vessel appears as the center line of the detection target. Then, the center line detection unit 36 sends an image in which the center line of the blood vessel appears as a result of the filtering process to the control unit 20 as the biometric identification data D6.

  When the control unit 20 receives the biometric identification data D6 from the image processing unit 22, the control unit 20 registers the identification data D6 as registration data in a recording medium (not shown) provided inside or outside, or exchanges the registration data with the registration data. The data is sent to a collator (not shown) provided inside or outside as a comparison target, and the control for the blood vessel imaging drive unit 21 and the image processing unit 22 is released.

  In this way, the information generating unit 3 can generate biometric identification data for identifying a biometric from the imaging result.

(4) Operation and Effect In the above configuration, the biological information generating apparatus 1 thickens the blood vessel included in the image obtained as a result of imaging the blood vessel existing in the living body, and thins the thickened blood vessel.

  Therefore, in this biometric information generation device 1, there is a discontinuity in the blood vessel in the image due to the small contrast difference between the blood vessel that is the identification target and the element corresponding to the noise other than the identification target. Even if it occurs, it can be thinned after joining the discontinuous parts, so regardless of the degree of other tissues intervening between the blood vessel and the biological surface, whether the imaging unit is good or bad, etc. Blood vessels can be accurately extracted.

  In addition, the biological information generation apparatus 1 binarizes an image obtained as an imaging result before the blood vessel is thickened. Therefore, in this biological information generating apparatus 1, blood vessels can be extracted accurately and at high speed even when a process with a large load is selected as the thickening process.

  Furthermore, this biological information generation apparatus 1 smoothes the image obtained as an imaging result before thickening the blood vessel. Therefore, since this biometric information generation device 1 can remove discontinuous points generated when edge extraction processing is performed on an image having a small contrast difference, noise that does not belong to a mesh-like identification target can be accurately removed. it can.

  For such smoothing, the first average value (FIG. 4B) in the peripheral pixel group of the target pixel NP (FIG. 4) in the image and the second average value in the target pixel NP and the peripheral pixel group are included. Depending on the difference from FIG. 4A, a method is applied in which the pixel value of the pixel of interest NP is replaced with the first average value, or the original value is not replaced.

  Therefore, in this biometric information generation device 1, as compared with the case where a smoothing process called a median filter is adopted, the processing load is reduced by the amount that does not require the sort process for detecting the intermediate value in the pixel of interest and the surrounding pixel group. Can smooth the image. Further, in this biological information generating apparatus 1, the smoothing processing result according to the present embodiment (FIG. 5A) and the smoothing that simply replaces the pixel value of the target pixel NP with the average value of the target pixel NP and the surrounding pixel group. As is apparent from a visual comparison of the processing (hereinafter referred to as simple smoothing processing) result (FIG. 5B), attention is paid as compared with the case of adopting the simple smoothing processing. The target pixel NP can be accurately smoothed without being affected by the noise component pixels included in the peripheral pixel group of the pixel NP.

  The experimental results are shown in FIGS. FIGS. 6 and 7 show changes in the pixel values of the central pixel of interest and the pixels adjacent to the pixel of interest. FIGS. 6A and 7A are diagrams before smoothing. 6 (B) and FIG. 7 (B) show the simple smoothing process after the smoothing process according to the present embodiment, and FIGS. 6 (C) and 7 (C) show the simple smoothing process, respectively. In FIGS. 6B and 7B, the replacement threshold employed in the smoothing process is set to “50”.

  As is clear from FIG. 6, the value of the target pixel is correctly replaced in the smoothing process according to the present embodiment, whereas the value of the target pixel NP is not correctly replaced in the simple smoothing process. . As is clear from FIG. 7, the smoothing process according to the present embodiment maintains the value of the target pixel without being affected by the noise component pixels included in the peripheral pixel group of the target pixel NP. In the simple smoothing process, it can be seen that the value of the pixel of interest is changed by being affected by the pixel of the noise component included in the peripheral pixel group.

  According to the above configuration, the blood vessel included in the image obtained as an imaging result of the blood vessel existing in the living body is thickened, and the thickened blood vessel is thinned. Even if a discontinuous portion occurs in the blood vessel in the image due to a small contrast difference with an element corresponding to noise other than the identification target, the thin line is obtained by combining the discontinuous portions. Therefore, it is possible to accurately extract blood vessels and thus extract blood vessels with high accuracy.

(5) Other Embodiments In the above-described embodiment, the case where a blood vessel inherent in a living body is applied as a net-like identification target in the living body has been described, but the present invention is not limited thereto, Various other things such as fingerprints that are present in the living body and nerves that are present in the living body can be applied. By the way, when a nerve is an authentication target, for example, if a marker specific to the nerve is injected into the body and the marker is imaged, the nerve is an identification target in the same manner as in the above-described embodiment. be able to.

  In addition, although the finger is applied as the range of the identification target, the present invention is not limited to this, and other biological parts such as an arm, a fundus, a palm, a toe, or a sole are applied. Alternatively, the whole living body may be applied.

  Further, as the imaging means for imaging the identification target, the configuration shown in FIG. 1 is applied. However, the present invention is not limited to this, and various configurations are possible depending on the application, type of identification target, and the like. Can be widely applied.

  In the above-described embodiment, the case where the filtering process called dilution is performed as the thickening means for thickening the identification target has been described. However, the present invention is not limited to this, and for example, a ones filter process. The identification target can be thickened by various other methods.

  Furthermore, in the above-described embodiment, the case where the filtering process called Thinning is performed as the thinning means for thinning the identification target has been described. However, the present invention is not limited to this, and for example, the blood vessel width is measured. Thus, the center line of the blood vessel may be detected.

  Furthermore, in the above-described embodiment, the case where the image processing unit 22 is operated in hardware has been described. However, the present invention is not limited to this, and the A / D conversion unit 31 and the edge in the image processing unit 22 are described. The extraction unit 32, the smoothing processing unit 33, the binarization unit 34, the thickening unit 35, and the centerline detection unit 36 may be operated in software by a program that executes all or part of the various processes.

  The present invention can be used in a field using a technique for identifying a living body.

It is a basic diagram which shows the whole structure of the biometric information generation apparatus by this Embodiment. It is a block diagram which shows the structure of an information generation part. It is a block diagram which shows the structure of an image process part. It is an approximate line figure used for explanation of smoothing processing. It is a basic diagram which shows the image of a smoothing process result. It is a basic diagram which shows an experimental result (1). It is a basic diagram which shows an experimental result (2).

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Living body information generation apparatus, 2 ... Blood vessel imaging part, 3 ... Information generation part, 14 ... Camera part, 15A, 15B ... Near infrared light source, 16 ... Macro lens, 17 ... Near red External light transmission filter, 18... CCD image sensor, 20... Control unit, 21... Blood vessel imaging drive unit, 22 ... image processing unit, 31... A / D conversion unit, 32. ...... Smoothing processing unit, 34... Binarization unit, 35... Thick line unit, 36 .. center line detection unit, COM .. command, S1 .. blood vessel projection image signal, D1. , D2 ... Blood vessel extraction image data, D3 ... Smoothed blood vessel extraction image data, D4 ... Blood vessel image data, D5 ... Thick blood vessel image data, D6 ... Biometric identification data, AR ... Setting range, NP ... Pixel of interest.

Claims (7)

Among the images obtained as a result of imaging a net-like identification target in a living body, thickening means for thickening the identification target;
An image processing apparatus comprising: thinning means for thinning the identification target thickened by the thickening means. The image processing apparatus according to claim 1, wherein the identification target is a blood vessel in a living body part. Binarizing means for binarizing an image obtained as an imaging result of an identification target in a living body,
The thickening means is
The image processing apparatus according to claim 1, wherein among the images binarized by the binarization unit, the identification target is bolded. Comprising a smoothing means for smoothing an image obtained as an imaging result of an identification target in a living body,
The smoothing means includes
According to the difference between the first average value in the peripheral pixel group of the target pixel in the image and the second average value in the target pixel and the peripheral pixel group, the pixel value of the target pixel is set to the first average value. Replace it with the value or leave it as it is,
The thickening means is
The image processing apparatus according to claim 1, wherein among the images smoothed by the smoothing unit, the identification target is thickened. Edge detection means for detecting an edge of an image obtained as a result of imaging a net-like identification target in a living body;
Smoothing means for smoothing an image whose edge is detected by the edge detection means;
Among the images smoothed by the smoothing means, thickening means for thickening the identification target;
An image processing apparatus comprising: thinning means for thinning the identification target thickened by the thickening means. A first step of thickening the identification target among images obtained as a result of imaging a net-shaped identification target in a living body;
And a second step of thinning the identification object that has been bolded. For a computer responsible for controlling an apparatus that processes an image obtained as an imaging result of an identification target in a living body,
A first process for thickening the identification target in the image;
And executing a second process of thinning the identification target that has been bolded.

Images