Identity recognition system and laser therapy device

Abstract

The present invention discloses an identity recognition system and a laser therapy device, which belongs to semiconductor laser therapy equipment; specifically comprises: a hand image acquisition unit, which is used to acquire the palm image of the user after straightening and the image after the hand is clenched into a fist; an image processing unit, which is connected to the hand image acquisition unit in communication, acquires the palm image and the image after the hand is clenched into a fist, compares them with the pre-stored user hand image information, and recognizes the user's identity. Through the rapid and accurate recognition of the identity, it can adapt to different user preference modes and improve the user's experience.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor laser therapy equipment, and in particular to an identity recognition system and a laser therapy device. Background Art

[0002] Laser is light generated by stimulated radiation of atoms. When the electrons in atoms absorb energy and jump from low energy level to high energy level, and then fall back from high energy level to low energy level, the released energy is released in the form of photons. When laser is irradiated on the skin, it has a certain degree of penetration. Laser energy of a specific wavelength can be selectively absorbed by hemoglobin. Oxygenated hemoglobin in blood vessels absorbs laser of corresponding wavelength and produces photothermal effect, which heats and coagulates diseased blood vessels. This is the selective photothermal effect. In addition, laser of a specific frequency can produce positive biomedical effects after irradiating human tissues, promote microcirculation of the human body, accelerate metabolism and the removal of harmful substances in blood vessels, and improve symptoms such as hyperlipidemia and blood viscosity by stimulating and improving the oxygen carrying capacity and deformability of red blood cells. It has a certain preventive and improvement effect on middle-aged and elderly people with vascular aging, especially patients with cardiovascular and cerebrovascular diseases. With the miniaturization and popularization of semiconductor lasers, the size of laser therapy equipment is getting smaller and smaller. Some wearable laser therapy devices have appeared on the market, which can perform laser therapy on specific parts of the human body and are suitable for home use.

[0003] However, different home users have different requirements for laser intensity and usage habits when using laser therapy devices. Different users need to frequently switch working modes and change preference settings, and cannot identify the user identity and adjust the output mode adaptively, resulting in a poor user experience.

[0004] Therefore, an identity recognition system and a laser therapy device are proposed, which can quickly identify the physiological characteristics of the user's identity, thereby determining the user's identity and adjusting the laser therapy output mode accordingly, without the need for frequent manual adjustment, which is very necessary for improving the user experience. Summary of the invention

[0005] In view of this, the present invention proposes an identity recognition system and a laser therapy device that perform identity recognition based on hand texture features, do not need to frequently remove a laser therapy device for identity verification, and have rapid information processing.

[0006] The technical solution of the present invention is implemented as follows: On the one hand, the present invention provides an identity recognition system, comprising:

[0007] A hand image acquisition unit, used to acquire an image of the user's straightened palm and an image of the user's hand after it is clenched into a fist;

[0008] The image processing unit is connected to the hand image acquisition unit for communication, and acquires the palm image and the image after the hand is clenched into a fist, and compares them with the pre-stored user's hand image information to identify the user's identity.

[0009] Based on the above technical solution, preferably, the workflow of the image processing unit is as follows:

[0010] S100, the hand image acquisition unit includes a voice prompt unit and a camera. After the user starts the hand image acquisition unit, the voice prompt unit prompts the user to stretch out his hand, keep the palm open and the five fingers together, so that the hand is completely within the field of view of the camera, and the current hand image is acquired and cropped, and the size of the palm image is saved according to a preset size; the image processing unit extracts the first feature and the second feature from the palm image;

[0011] S200, the voice prompting unit further prompts the user to clench the hand, curl up the four fingers of the hand except the thumb toward the palm, keep the thumb straight and do not cover the four fingers of the hand except the thumb, obtain the current hand image for cropping, and save the size of the image after the hand is clenched into a fist according to a preset size; the image processing unit extracts the third feature from the image after the hand is clenched into a fist;

[0012] S300. If the image processing unit does not pre-store the palm image, the image after the hand is clenched into a fist, and the corresponding first feature, second feature, and third feature, the palm image and the image after the hand is clenched into a fist acquired this time are stored as the default user hand image information, and the corresponding first feature, second feature, and third feature are extracted.

[0013] Preferably, the current hand image is acquired and cropped, and the size of the palm image is saved according to a preset size; the image processing unit extracts the first feature and the second feature from the palm image, and the specific content is:

[0014] S101, the camera obtains an initial image of the environment;

[0015] S102, the voice prompting unit broadcasts a message to prompt the user to bend the arm, with the forearm bent toward the head and neck, so that the open palm faces the camera's field of view, and obtains an initial palm image;

[0016] S103, aligning the initial palm image and the initial environment image and performing Boolean operations to remove the background content to obtain a palm image; calculating the grayscale of each pixel of the palm image by weighted average to obtain a grayscale image: Gary = 0.299R + 0.587G + 0.114B, where Gary is the grayscale value of the pixel of the palm image, and R, G and B are the red pixel value, green pixel value and blue pixel value of the pixel of the palm image respectively;

[0017] S104, defining the horizontal extension direction of the middle finger as the x-axis direction, and the direction perpendicular to the horizontal extension direction of the middle finger as the y-axis direction, rotating the grayscale image so that the x-axis direction is consistent with the length direction of the palm image, and the y-axis direction is consistent with the height direction of the palm image; extracting the palm print in the grayscale image to obtain an extracted palm print texture image;

[0018] S105, performing curve fitting on the palmprint texture images in different directions, retaining N palmprint texture curves, marking the bifurcation points on the same palmprint texture curve or the intersection points of different palmprint texture curves; taking the N palmprint texture curves as the first features, and the bifurcation points and intersection points on the palmprint texture curves as the second features.

[0019] Further preferably, the step S104 of extracting the palm print from the grayscale image to obtain the extracted palm print texture image is to extract the palm print texture from the grayscale image using a Gabor filter. The expression is Where σ is the standard deviation of the Gaussian function used by the Gabor filter; λ is the wavelength of the cosine function part; is the phase shift of the cosine function part; γ is the spatial aspect ratio, which indicates the ellipticity of the Gabor filter; θ is the direction, x′=-x cosθ+y sinθ, y′=-x sinθ+y cosθ, where x and y are the coordinates of the pixel point in the grayscale image, respectively, and x′ and y′ are the coordinates of the pixel point (x, y) after rotation; the response image R(x, y) in the direction θ is obtained as follows: R(x, y)=I(x, y)*g(x, y; λ, θ, ψ, σ, γ), where I(x, y) is the grayscale image, * is the convolution operation, and the response image in the direction θ is obtained by the operation, and the response images in different directions are combined to obtain the extracted palmprint texture image.

[0020] More preferably, the specific content of step S105 is: performing segmented curve fitting on the extracted palmprint texture images respectively, and the segmented curve fitting satisfies the following rules: 1) the same palmprint texture image is segmented, and each segmented curve includes at least three adjacent bifurcation points and / or intersection points; 2) each segmented curve is fitted respectively; 3) different bifurcation points or intersection points of two adjacent segmented curves are collinear, and have second-order continuity in the neighborhood of the common endpoint; 4) the first N palmprint texture images containing the largest number of pixels are retained, and the angle range between the beginning and the end of each palmprint texture image is (90°, 180°).

[0021] Preferably, the specific content of step S200 is:

[0022] S201, keep the finger joints clean and dry, keep the thumb straight, and the voice prompt unit broadcasts to prompt the user to maintain the current position of the palm and make a fist, and make a fist with four fingers outside the thumb; make the palm of the fist face the camera's field of view to obtain an initial fist image;

[0023] S202, aligning the initial fist image with the initial environment image and performing Boolean operations to remove background content to obtain an image of the hand after making a fist;

[0024] S203, adjusting the contrast and brightness of the image after the hand is clenched into a fist, and enhancing the clarity of the texture on the back of the fingers;

[0025] S204, rotating the image after the hand is clenched into a fist with adjusted contrast and brightness, so that the direction in which the middle finger extends is parallel to the x-axis direction, and the height direction of the image after the hand is clenched into a fist is consistent with the y-axis; creating a picking window for each joint of the bent finger, respectively, to obtain a local image of the first joint at the fingertips of the four bent fingers;

[0026] S205: extract the vertical texture of the back of the finger from the partial image, and use the vertical texture as the third feature.

[0027] Further preferably, the creation of the picking window in step S204 is to use the longitudinal tangent of the edge of the nail as the starting position, offset the first pixel length and the second pixel length from the starting position along the positive semi-axis direction of the x-axis respectively, to obtain the first boundary and the second boundary, and connect the first boundary, the second boundary and the finger edge area to form a closed picking window; the distance of the real world coordinate corresponding to the offset of the first pixel length is determined according to the following relationship: Where a1 is the age factor, which is the real age; a2 is the height factor, in millimeters; a3 is the gender factor, which is 2 for males and 1 for females; the distance of the real world coordinate corresponding to the offset of the second pixel length is determined according to the following relationship: Through the transformation relationship between the world coordinate system and the pixel coordinate system of the camera, the distance in the world coordinate system is converted into the first pixel length and the second pixel length in the pixel coordinate system; the finger edge area is determined by the change of the cumulative value of the RGB pixel value in the local image.

[0028] Further preferably, the vertical texture on the back of the finger extracted from the local image in step S205 takes the vertical texture as the third feature, offsets a certain number of pixels inward from the boundary of the picking window, fits the vertical texture within the offset boundary to determine the image of the vertical texture, and takes the number of vertical textures and the pixel length of the vertical texture as the third feature.

[0029] Still further preferably, the image processing unit obtains the image information corresponding to the hand, compares it with the pre-stored image information of the user's hand, and identifies the identity of the user by performing the following identity comparison process:

[0030] Step 1) The user starts the hand image acquisition unit;

[0031] Step 2) The hand image acquisition unit prompts the user to bend his arm, and the camera acquires the palm image; then prompts the user to clench his hand to obtain an image after the hand is clenched into a fist; the current user's palm image and the image after the hand is clenched into a fist are adjusted in size and direction according to a preset size;

[0032] Step 3) extracting a palm print texture image from the palm image to obtain a first feature and a second feature of the current user's hand; extracting a vertical texture of the back of the fingers from the image after the hand is clenched into a fist to obtain a third feature corresponding to the vertical texture of the back of the fingers of the current user;

[0033] Step 4) Select a number of bifurcation points or intersection points in the second feature of the current user's hand, and determine whether no less than 85% of the total number of bifurcation points or intersection points are located in the neighborhood range of the palm print texture curve in the user's hand image information pre-stored in the image processing unit. If so, it is considered that the first feature and the second feature of the current user are similar to the first feature and the second feature pre-stored in the image processing unit, and the current user is a saved historical user; if not, further execute step 5);

[0034] Step 5) obtain the third feature corresponding to the vertical texture of the back of the current user's finger, offset a certain number of pixels inward at the boundary of the picking window to form a virtual boundary, construct a number of first virtual straight lines parallel to the x-axis within the virtual boundary, set the pixel interval of adjacent first virtual straight lines, and count the number of first intersections between the first virtual straight lines and the vertical texture of the back of the current user's finger; construct a number of second virtual straight lines using virtual boundaries of the same size in the vertical texture corresponding to the third feature pre-stored in the image processing unit, the second virtual straight lines are parallel to the x-axis and have the same pixel interval, count the number of second intersections between the second virtual straight line and the vertical texture corresponding to the third feature pre-stored, compare the vertical texture of the back of the current user's finger, the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections; if the number of vertical textures on the back of the current user's finger is different from the number of vertical textures on the back of the image processing unit, then the vertical texture of the back of the current user is compared with the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections When the number of vertical textures in the third feature pre-stored in the processing unit is the same, and the difference between the first number of intersections and the second number of intersections does not exceed the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of the vertical texture on the back of the finger of the current user and the pixel length of the vertical texture in the third feature pre-stored in the image processing unit are counted in order from small to large in the x-axis and y-axis coordinates. If the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions does not exceed 5%, the current user is considered to be a saved historical user; if the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions exceeds 5%, the difference between the first number of intersections and the second number of intersections exceeds the number of vertical textures in the third feature pre-stored in the image processing unit, or the number of vertical textures on the back of the finger of the current user is different from the number of vertical textures in the third feature pre-stored in the image processing unit, jump to step 6);

[0035] Step 6) If the current user is identified as a saved historical user in either step 4) or step 5), the identity and usage preferences corresponding to the historical user are called to perform laser therapy; if neither step 4) nor step 5) identifies the current user as a saved historical user, the current user is considered to be a new user, and the image processing unit establishes a dedicated storage space for the first feature, the second feature, and the third feature for the current user and numbers the new user, and saves the parameter settings of the new user so that it can be directly called for laser therapy the next time it is used.

[0036] On the other hand, the present application provides a laser therapy device, including the above-mentioned identity recognition system, as well as a main body and an irradiation unit; the main body is worn on the user's neck; the hand image acquisition unit is arranged on the outer surface of the main body away from the neck; the irradiation unit is embedded in the surface of the main body close to the neck, for laser irradiation of different positions of the neck; the image processing unit is arranged inside the main body, and the image processing unit is also communicated with the irradiation unit for driving the irradiation unit into a working state.

[0037] The identity recognition system and laser therapy device provided by the present invention have the following beneficial effects compared with the prior art:

[0038] The present application adopts the method of wearing the device first and then verifying the identity, and selects palm prints and vertical lines on the back of fingers with obvious individual differences as feature sources for identifying the identity. Through image segmentation, image preprocessing and texture extraction, curves and feature points are quickly obtained as comparison content, and the palm and back of the hand images are quickly acquired and verified twice. Only one image needs to pass the verification to determine that the current user is a historical user. This can greatly improve the convenience of using the laser therapy device, and greatly reduce the amount of calculation and processing time compared to the facial recognition method. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 This is a framework diagram of an identity recognition system and a laser therapy device of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] like Figure 1 As shown, on the one hand, the present invention provides an identity recognition system, comprising:

[0043] The hand image acquisition unit is used to acquire image information corresponding to the hand; the hand image acquisition unit is used to acquire images of the palm prints and the fingertip joints on the back of the hand.

[0044] The image processing unit is connected to the hand image acquisition unit for communication, obtains the image information corresponding to the hand, compares it with the pre-stored user's hand image information, and identifies the user's identity. After receiving the palm prints and the images of the fingertips on the back of the hand, the image processing unit extracts features as a basis for identity recognition.

[0045] Palm prints refer to the lines and patterns on the surface of the skin on the palm, including dermatoglyphics, sweat pores, and bumps on the skin. Palm prints are unique to each person and can be used for identity recognition. Horizontal lines on the back of fingers refer to the horizontal lines and patterns on the skin on the back of fingers. Similar to palm prints, horizontal lines on the back of fingers are also unique biometric features of each person and can be used for identity recognition. Compared with traditional facial recognition, which requires long-term immobility, multiple facial features, large amount of calculation, high calculation delay, and easy failure of image acquisition, the combination of palm prints and horizontal lines on the back of fingers is a way to improve the accuracy of identity recognition and reduce the amount of recognition calculation and delay.

[0046] In one embodiment, the image processing unit has the following workflow:

[0047] S100, the hand image acquisition unit includes a voice prompt unit and a camera. After the user starts the hand image acquisition unit, the voice prompt unit prompts the user to stretch out his hand, keep the palm open and the five fingers together, so that the hand is completely within the field of view of the camera, and the current hand image is acquired and cropped, and the size of the palm image is saved according to a preset size; the image processing unit extracts the first feature and the second feature from the palm image.

[0048] Specifically, step S100 includes the following sub-steps:

[0049] S101, the camera obtains an initial environment image; the initial environment image reflects the background image of the environment where the user is located, and is usually a static environment image.

[0050] S102, the voice prompting unit broadcasts, prompting the user to bend the arm, with the forearm bent toward the head and neck, so that the open palm faces the camera's field of view, and obtains an initial palm image; the initial palm image includes a background image and an image of the palm part of the foreground.

[0051] S103, aligning the initial palm image and the initial environment image and performing Boolean operation to remove the background content to obtain the palm image; calculating the grayscale of each pixel of the palm image by weighted average to obtain a grayscale image: Gary = 0.299R + 0.5887G + 0.114B, Gary is the grayscale value of the pixel of the palm image, R, G and B are the red pixel value, green pixel value and blue pixel value of the pixel of the palm image respectively; in this process, a reference object in the initial environment image can be selected as a reference, and the initial palm image and the initial environment image are aligned by rotation or translation operation, and the same background content in the initial palm image and the initial environment image is removed by sampling Boolean subtraction operation to obtain a palm image containing only the palm content. Then, the palm image is converted from the RGB color space to a more reasonable grayscale space in order to simplify the dimension of the image data, adjust the morphological characteristics of the image, and improve the subsequent operation speed. The weighted average method calculates the grayscale of the pixel points. Based on the preference of the human eye, the weight of the red pixel value is 0.299, the weight of the green pixel value is 0.587, and the weight of the blue pixel value is 0.114. Each weight is determined based on the sensitivity of the human eye to different colors. This method retains the brightness information of the palm image. The above weights are empirical values ​​and can be adjusted according to actual needs. I will not go into details here.

[0052] S104, define the horizontal extension direction of the middle finger as the x-axis direction, and the direction perpendicular to the horizontal extension direction of the middle finger as the y-axis direction, rotate the grayscale image so that the x-axis direction is consistent with the length direction of the palm image, and the y-axis direction is consistent with the height direction of the palm image; extract the palm print in the grayscale image to obtain the extracted palm print texture image.

[0053] In this embodiment, the palm print is extracted by using a Gabor filter for directional filtering to obtain textures in different directions. The expression is Where σ is the standard deviation of the Gaussian function used by the Gabor filter; λ is the wavelength of the cosine function part; is the phase shift of the cosine function part; γ is the spatial aspect ratio, which indicates the ellipticity of the Gabor filter; θ is the direction, x′=x cosθ+y sinθ, y′=-x sinθ+y cosθ, where x and y are the coordinates of the pixel point in the grayscale image, respectively, and x′ and y′ are the coordinates of the pixel point (x, y) after rotation; the response image R(x, y) in the direction θ is obtained as follows: R(x, y)=I(x, y)*g(x, y; λ, θ, ψ, σ, γ), where I(x, y) is the grayscale image, * is the convolution operation, and the response image in the direction θ is obtained by the operation, and the response images in different directions are combined to obtain the extracted palmprint texture image. The Gabor filter can only be stretched along the x-axis or y-axis of the grayscale image I (x, y), but not diagonally, so the direction θ is introduced. After the direction θ is introduced, the original coordinates (x, y) are converted to (x′, y′), which allows the Gabor filter to be stretched along the direction θ, thereby obtaining images with responses in different directions. The window shape of the Gabor filter can be a rectangular window or a square window, preferably a square window with odd-valued pixels.

[0054] S105, performing curve fitting on the palmprint texture images in different directions, retaining N palmprint texture curves, marking the bifurcation points on the same palmprint texture curve or the intersection points of different palmprint texture curves; taking the N palmprint texture curves as the first features, and the bifurcation points and intersection points on the palmprint texture curves as the second features.

[0055] The above-mentioned segmented curve fitting meets the following rules: 1) the same palm print texture image is segmented, and each segmented curve includes at least three adjacent bifurcation points and / or intersection points; 2) each segmented curve is fitted separately; 3) the different bifurcation points or intersection points of two adjacent segmented curves are collinear, and have second-order continuity in the neighborhood of the common endpoint; 4) the first N palm print texture images containing the largest number of pixels are retained, and the angle range between the beginning and the end of each palm print texture image is (90°, 180°). The segmented curve fitting here can adopt Bezier curve fitting or B curve fitting. In this embodiment, Bezier curve fitting is preferably used for curve fitting. N is an odd number not less than 3.

[0056] The reason why the segmented fitting method is adopted is that the palmprint texture curve does not change linearly. Through segmentation, it can be better expressed and fitted, reducing the difficulty of overall fitting, and limiting the relative position of the second feature in the palmprint texture curve. Limiting the angle range of the palmprint texture image is to avoid the interference caused by the extraction of bifurcation and to eliminate palmprints that are too short and discontinuous. Since the direction of the grayscale image is limited, the starting point of each palmprint texture curve and the pixel coordinate position of the second feature in the grayscale image are also determined, which is convenient for the subsequent comparison process.

[0057] S200. After completing the extraction of the palm image, the voice prompt unit further prompts the user to clench the hand, curl up the four fingers of the hand except the thumb toward the palm, keep the thumb straight and do not cover the four fingers of the hand except the thumb, obtain the current hand image for cropping, and save the size of the image after the hand is clenched into a fist according to the preset size; the image processing unit extracts the third feature from the image after the hand is clenched into a fist.

[0058] The specific content of step S200 is:

[0059] S201. Keep the finger joints clean and dry, and keep the thumb straight. The voice prompt unit broadcasts to prompt the user to maintain the current position of the palm and make a fist with the four fingers outside the thumb; make the palm of the fist face the camera's field of view to obtain the initial image of the fist; since the initial image of the environment has been extracted before, there is no need to extract it again here.

[0060] S202, similarly performing a sampling Boolean subtraction operation, eliminating the same background content in the fist initial image and the environment initial image, and obtaining a palm image containing only the palm content.

[0061] S203, adjusting the contrast and brightness of the image after the hand is clenched into a fist to enhance the clarity of the texture on the back of the finger. Before adjusting the brightness and contrast, the image after the hand is clenched into a fist is converted into a grayscale image by weighted averaging. This operation of reducing the brightness and increasing the contrast can improve the clarity of the image and facilitate the subsequent extraction of the vertical texture on the back of the finger.

[0062] Brightness enhancement is achieved in the following ways: new =L old +b, where L old is the original pixel value, b is the enhanced brightness value, L new is the pixel value after brightness enhancement. Contrast enhancement is performed in the following way: new =a×D old , where D old is the original pixel value, a is the contrast enhancement coefficient, D new is the pixel value after contrast enhancement.

[0063] S204. Rotate the image of the clenched fist with adjusted contrast and brightness so that the direction in which the middle finger extends is parallel to the x-axis and the height direction of the image of the clenched fist is consistent with the y-axis; create a picking window for each joint of the bent finger to obtain a local image of the first joint at the fingertips of the four bent fingers.

[0064] The key point of this step is how to create a picking window. It is necessary to determine the area range for filtering vertical textures. This embodiment adopts the method of constructing a first boundary and a second boundary, and then combining the finger edge area to form a picking window. Specifically:

[0065] First, the longitudinal tangent of the edge of the nail is used as the starting position, and the first pixel length and the second pixel length are offset from the starting position along the positive semi-axis direction of the x-axis respectively to obtain the first boundary and the second boundary; then the pixels corresponding to the finger edge area are obtained, and the first boundary, the second boundary and the finger edge area are connected to form a closed picking window.

[0066] Here, the offset first pixel length and the offset second length are different for users of different ages, heights or genders. First, the distance between the first knuckle range and the nail edge of people of different ages and heights is fitted, and then the distance is converted to the pixel coordinate system and converted into pixel distance. The distance of the real world coordinate corresponding to the offset first pixel length is determined according to the following relationship: Where a1 is the age factor, which is the real age; a2 is the height factor, in millimeters; a3 is the gender factor, which is 2 for males and 1 for females; the distance of the real world coordinate corresponding to the offset of the second pixel length is determined according to the following relationship: Through the transformation relationship between the world coordinate system and the pixel coordinate system of the camera, the distance in the world coordinate system is converted into the first pixel length and the second pixel length in the pixel coordinate system. The first knuckle part is selected because the number of vertical textures in this part is small and easy to identify, which can improve the speed of subsequent image comparison and reduce time delay.

[0067] The finger edge area is determined by the change in the cumulative value of the RGB pixel value in the local image. Pixels that meet the following conditions are considered to be pixels in the finger edge area: (1) Calculate the cumulative value of the RGB pixel value of each pixel in the positive semi-axis direction of the x-axis. If the cumulative value of the RGB pixel value of the current pixel does not exceed 70% of the cumulative value of the RGB pixel value of the adjacent pixel in the y-axis direction, it is considered to be a potential boundary pixel; (2) The tangent direction of the adjacent potential boundary pixel points is from the first boundary to the second boundary, and the angle between the tangent direction and the first boundary or the second boundary does not exceed 90°. By default, the continuous dark part is taken as the boundary of the four bent fingers. By comparing with the adjacent pixels, can the pixels of the potential finger edge area be obtained? By further confirming the extension direction of the pixels and filtering out the dark vertical texture, the pixels of the finger edge area can be obtained. The pixels of the finger edge are fitted into a straight line or curve, connected with the first boundary and the second boundary to form a closed picking window.

[0068] S205: extract the vertical texture of the back of the finger from the partial image, and use the vertical texture as the third feature.

[0069] Specifically, the boundary of the picking window is offset inward by a certain number of pixels, and the vertical texture is fitted within the offset boundary to determine the image of the vertical texture, and the number of vertical textures and the pixel length of the vertical texture are used as the third feature. Since the color of the pixels at the edge of the finger is also relatively dark, in order to prevent mutual interference, a certain amount of pixel offset is performed, and the pixel offset amount is not less than 3 times the pixel width of the edge of the finger.

[0070] S300. If the image processing unit does not pre-store the palm image, the image after the hand is clenched into a fist, and the corresponding first feature, second feature, and third feature, the palm image and the image after the hand is clenched into a fist acquired this time are stored as the default user hand image information, and the corresponding first feature, second feature, and third feature are extracted.

[0071] The image processing unit performs identity comparison as follows:

[0072] Step 1) The user starts the hand image acquisition unit;

[0073] Step 2) The hand image acquisition unit prompts the user to bend his arm, and the camera acquires the palm image; then prompts the user to clench his hand to obtain an image after the hand is clenched into a fist; the current user's palm image and the image after the hand is clenched into a fist are adjusted in size and direction according to a preset size;

[0074] Step 3) extracting a palm print texture image from the palm image to obtain a first feature and a second feature of the current user's hand; extracting a vertical texture of the back of the fingers from the image after the hand is clenched into a fist to obtain a third feature corresponding to the vertical texture of the back of the fingers of the current user;

[0075] Step 4) Select a number of bifurcation points or intersection points in the second feature of the current user's hand, and determine whether no less than 85% of the total number of bifurcation points or intersection points are located in the neighborhood range of the palm print texture curve in the user's hand image information pre-stored in the image processing unit. If so, it is considered that the first feature and the second feature of the current user are similar to the first feature and the second feature pre-stored in the image processing unit, and the current user is a saved historical user; if not, further execute step 5); in this embodiment, the neighborhood range of the palm print texture curve is set to a deviation of 5 pixels on both sides of the palm print texture curve.

[0076] Step 5) obtain the third feature corresponding to the vertical texture of the back of the current user's finger, offset a certain number of pixels inward at the boundary of the picking window to form a virtual boundary, construct a number of first virtual straight lines parallel to the x-axis within the virtual boundary, set the pixel interval of adjacent first virtual straight lines, and count the number of first intersections between the first virtual straight lines and the vertical texture of the back of the current user's finger; construct a number of second virtual straight lines using virtual boundaries of the same size in the vertical texture corresponding to the third feature pre-stored in the image processing unit, the second virtual straight lines are parallel to the x-axis and have the same pixel interval, count the number of second intersections between the second virtual straight line and the vertical texture corresponding to the third feature pre-stored, compare the vertical texture of the back of the current user's finger, the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections; if the number of vertical textures on the back of the current user's finger is different from the number of vertical textures on the back of the image processing unit, then the vertical texture of the back of the current user is compared with the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections When the number of vertical textures in the third feature pre-stored in the processing unit is the same, and the difference between the first number of intersections and the second number of intersections does not exceed the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of the vertical texture on the back of the finger of the current user and the pixel length of the vertical texture in the third feature pre-stored in the image processing unit are counted in order from small to large in the x-axis and y-axis coordinates. If the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions does not exceed 5%, the current user is considered to be a saved historical user; if the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions exceeds 5%, the difference between the first number of intersections and the second number of intersections exceeds the number of vertical textures in the third feature pre-stored in the image processing unit, or the number of vertical textures on the back of the finger of the current user is different from the number of vertical textures in the third feature pre-stored in the image processing unit, jump to step 6);

[0077] The process first confirms whether the number of vertical textures of the currently acquired image is the same as the number of vertical textures pre-stored in the image processing unit, and then compares the first number of intersections with the second number of intersections. If the difference in the number of intersections does not exceed the number of vertical textures, it means that further comparison can be made, that is, comparing the pixel length of each attribute texture, that is, the number of pixels. If the difference in pixel length is within an acceptable range, it is considered that the current user is a historical user.

[0078] Step 6) If the current user is identified as a saved historical user in either step 4) or step 5), the identity and usage preferences corresponding to the historical user are called to perform laser therapy; if neither step 4) nor step 5) identifies the current user as a saved historical user, the current user is considered to be a new user, and the image processing unit establishes a dedicated storage space for the first feature, the second feature, and the third feature for the current user and numbers the new user, and saves the parameter settings of the new user so that it can be directly called for laser therapy the next time it is used.

[0079] In addition, the present application also provides a laser therapy device, including the above-mentioned identity recognition system and an irradiation unit; the main body is worn on the neck of the user; the hand image acquisition unit is arranged on the outer surface of the main body away from the neck; the irradiation unit is embedded in the surface of the main body close to the neck, and is used to perform laser irradiation on different positions of the neck; the image processing unit is arranged inside the main body, and the image processing unit is also communicated with the irradiation unit to drive the irradiation unit into a working state. The main body can be provided with multiple irradiation units, each of which has a different number of semiconductor lasers arranged in an array, which are used to output lasers to different parts of the neck. After the image processing unit obtains the first feature, the second feature and the third feature of the user, it can easily confirm the identity of the user when it is used next time, and select different laser output modes according to the user's preference.

[0080] Specifically, the hand image acquisition unit includes a voice prompt unit and a camera, both of which are arranged on the side of the body away from the surface of the user's neck, and the lens of the camera faces the front area of ​​the neck; the voice prompt unit is used to issue a voice prompt after the user wears the body, so as to let the user stretch out his hand, and the camera obtains the image of the user's palm after straightening and the image of the user's palm after clenching a fist.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An identity recognition system, characterized in that: include: A hand image acquisition unit, used to acquire an image of the user's straightened palm and an image of the user's hand after it is clenched into a fist; The image processing unit is connected to the hand image acquisition unit for communication, and acquires the palm image and the image after the hand is clenched into a fist, and compares them with the pre-stored user's hand image information to identify the user's identity; The workflow of the image processing unit is as follows: S100, the hand image acquisition unit includes a voice prompt unit and a camera. After the user starts the hand image acquisition unit, the voice prompt unit prompts the user to stretch out his hand, keep the palm open and the five fingers together, so that the hand is completely within the field of view of the camera, and the current hand image is acquired and cropped, and the size of the palm image is saved according to a preset size; the image processing unit extracts the first feature and the second feature from the palm image; S200, the voice prompting unit further prompts the user to clench the hand, curl up the four fingers of the hand except the thumb toward the palm, keep the thumb straight and do not cover the four fingers of the hand except the thumb, obtain the current hand image for cropping, and save the size of the image after the hand is clenched into a fist according to a preset size; the image processing unit extracts the third feature from the image after the hand is clenched into a fist; S300: If the image processing unit does not pre-store the palm image, the image after the hand is clenched into a fist, and the corresponding first feature, second feature, and third feature, the palm image and the image after the hand is clenched into a fist acquired this time are stored as default user hand image information, and the corresponding first feature, second feature, and third feature are extracted; The current hand image is obtained and cropped, and the size of the palm image is saved according to a preset size; the image processing unit extracts the first feature and the second feature from the palm image, and the specific content is: S101, the camera obtains an initial image of the environment; S102, the voice prompting unit broadcasts a message to prompt the user to bend the arm, with the forearm bent toward the head and neck, so that the open palm faces the camera's field of view, and obtains an initial palm image; S103, aligning the initial palm image and the initial environment image and performing Boolean operations to remove the background content to obtain a palm image; calculating the grayscale of each pixel of the palm image by weighted average to obtain a grayscale image: Gary = 0.299R + 0.587G + 0.114B, where Gary is the grayscale value of the pixel of the palm image, and R, G and B are the red pixel value, green pixel value and blue pixel value of the pixel of the palm image respectively; S104, defining the horizontal extension direction of the middle finger as the x-axis direction, and the direction perpendicular to the horizontal extension direction of the middle finger as the y-axis direction, rotating the grayscale image so that the x-axis direction is consistent with the length direction of the palm image, and the y-axis direction is consistent with the height direction of the palm image; extracting the palm print in the grayscale image to obtain an extracted palm print texture image; S105, performing curve fitting on palmprint texture images in different directions, retaining N palmprint texture curves, marking bifurcation points on the same palmprint texture curve or intersection points of different palmprint texture curves; taking the N palmprint texture curves as first features, and taking the bifurcation points and intersection points on the palmprint texture curves as second features; The specific content of step S200 is: S201, keep the finger joints clean and dry, keep the thumb straight, and the voice prompt unit broadcasts to prompt the user to maintain the current position of the palm and make a fist, and make a fist with four fingers outside the thumb; make the palm of the fist face the camera's field of view to obtain an initial fist image; S202, aligning the initial fist image with the initial environment image and performing Boolean operations to remove background content to obtain an image of the hand after making a fist; S203, adjusting the contrast and brightness of the image after the hand is clenched into a fist, and enhancing the clarity of the texture on the back of the fingers; S204, rotating the image after the hand is clenched into a fist with adjusted contrast and brightness, so that the direction in which the middle finger extends is parallel to the x-axis direction, and the height direction of the image after the hand is clenched into a fist is consistent with the y-axis; creating a picking window for each joint of the bent finger, respectively, to obtain a local image of the first joint at the fingertips of the four bent fingers; S205: extract the vertical texture of the back of the finger from the partial image, and use the vertical texture as the third feature.

2. An identity recognition system according to claim 1, characterized in that: The step S104 of extracting the palm print from the grayscale image to obtain the extracted palm print texture image is to extract the palm print texture from the grayscale image using a Gabor filter. The expression is Where σ is the standard deviation of the Gaussian function used by the Gabor filter; λ is the wavelength of the cosine function part; is the phase shift of the cosine function part; γ is the spatial aspect ratio, which indicates the ellipticity of the Gabor filter; θ is the direction, x′=x cosθ+y sinθ, y′=-x sinθ+y cosθ, where x and y are the coordinates of the pixel point in the grayscale image, respectively, and x' and y' are the coordinates of the pixel point (x, y) after rotation; the response image R(x, y) in the direction θ is obtained as follows: R(x, y)=I(x, y)*g(x, y; λ, θ, ψ, σ, γ), where I(x, y) is the grayscale image, * is the convolution operation, and the response image in the direction θ is obtained by the operation, and the response images in different directions are combined to obtain the extracted palmprint texture image.

3. An identity recognition system according to claim 2, characterized in that: The specific content of step S105 is: performing segmented curve fitting on the extracted palmprint texture images respectively, and the segmented curve fitting meets the following rules: 1) the same palmprint texture image is segmented, and each segmented curve includes at least three adjacent bifurcation points and / or intersection points; 2) each segmented curve is fitted respectively; 3) different bifurcation points or intersection points of two adjacent segmented curves are collinear, and have second-order continuity in the neighborhood of the common endpoint; 4) the first N palmprint texture images containing the largest number of pixels are retained, and the angle range between the beginning and the end of each palmprint texture image is (90°, 180°).

4. An identity recognition system according to claim 1, characterized in that: The creation of the picking window described in step S204 is to use the longitudinal tangent of the edge of the nail as the starting position, offset the first pixel length and the second pixel length from the starting position along the positive semi-axis direction of the x-axis respectively, to obtain the first boundary and the second boundary, and connect the first boundary, the second boundary and the finger edge area to form a closed picking window; the distance of the real world coordinate corresponding to the offset of the first pixel length is determined according to the following relationship: Where a1 is the age factor, which is the real age; a2 is the height factor, in millimeters; a3 is the gender factor, which is 2 for males and 1 for females; the distance of the real world coordinate corresponding to the offset of the second pixel length is determined according to the following relationship: Through the transformation relationship between the world coordinate system and the pixel coordinate system of the camera, the distance in the world coordinate system is converted into the first pixel length and the second pixel length in the pixel coordinate system; the finger edge area is determined by the change of the cumulative value of the RGB pixel value in the local image.

5. An identity recognition system according to claim 4, characterized in that: The vertical texture of the back of the finger extracted from the local image in step S205 takes the vertical texture as the third feature, offsets a certain number of pixels inward from the boundary of the picking window, fits the vertical texture within the offset boundary to determine the image of the vertical texture, and takes the number of vertical textures and the pixel length of the vertical texture as the third feature.

6. An identity recognition system according to claim 5, characterized in that: The image processing unit obtains the image information corresponding to the hand, compares it with the pre-stored image information of the user's hand, and identifies the user's identity, which is the following identity comparison process: Step 1) The user starts the hand image acquisition unit; Step 2) The hand image acquisition unit prompts the user to bend his arm, and the camera acquires the palm image; then prompts the user to clench his hand to obtain an image after the hand is clenched into a fist; the current user's palm image and the image after the hand is clenched into a fist are adjusted in size and direction according to a preset size; Step 3) extracting a palm print texture image from the palm image to obtain a first feature and a second feature of the current user's hand; extracting a vertical texture of the back of the fingers from the image after the hand is clenched into a fist to obtain a third feature corresponding to the vertical texture of the back of the fingers of the current user; Step 4) Select a number of bifurcation points or intersection points in the second feature of the current user's hand, and determine whether no less than 85% of the total number of bifurcation points or intersection points are located in the neighborhood range of the palm print texture curve in the user's hand image information pre-stored in the image processing unit. If so, it is considered that the first feature and the second feature of the current user are similar to the first feature and the second feature pre-stored in the image processing unit, and the current user is a saved historical user; if not, further execute step 5); Step 5) obtain the third feature corresponding to the vertical texture of the back of the current user's finger, offset a certain number of pixels inward at the boundary of the picking window to form a virtual boundary, construct a number of first virtual straight lines parallel to the x-axis within the virtual boundary, set the pixel interval of adjacent first virtual straight lines, and count the number of first intersections between the first virtual straight lines and the vertical texture of the back of the current user's finger; construct a number of second virtual straight lines using virtual boundaries of the same size in the vertical texture corresponding to the third feature pre-stored in the image processing unit, the second virtual straight lines are parallel to the x-axis and have the same pixel interval, count the number of second intersections between the second virtual straight line and the vertical texture corresponding to the third feature pre-stored, compare the vertical texture of the back of the current user's finger, the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections; if the number of vertical textures on the back of the current user's finger is different from the number of vertical textures on the back of the image processing unit, then the vertical texture of the back of the current user is compared with the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of each vertical texture, and the difference between the first number of intersections and the second number of intersections When the number of vertical textures in the third feature pre-stored in the processing unit is the same, and the difference between the first number of intersections and the second number of intersections does not exceed the number of vertical textures in the third feature pre-stored in the image processing unit, the pixel length of the vertical texture on the back of the finger of the current user and the pixel length of the vertical texture in the third feature pre-stored in the image processing unit are counted in order from small to large in the x-axis and y-axis coordinates. If the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions does not exceed 5%, the current user is considered to be a saved historical user; if the difference in pixel lengths of at least three vertical textures with adjacent endpoint positions exceeds 5%, the difference between the first number of intersections and the second number of intersections exceeds the number of vertical textures in the third feature pre-stored in the image processing unit, or the number of vertical textures on the back of the finger of the current user is different from the number of vertical textures in the third feature pre-stored in the image processing unit, jump to step 6); Step 6) If the current user is identified as a saved historical user in either step 4) or step 5), the identity and usage preferences corresponding to the historical user are called to perform laser therapy; if neither step 4) nor step 5) identifies the current user as a saved historical user, the current user is considered to be a new user, and the image processing unit establishes a dedicated storage space for the first feature, the second feature, and the third feature for the current user and numbers the new user, and saves the parameter settings of the new user so that it can be directly called for laser therapy the next time it is used.

7. A laser therapy device, characterized in that: It comprises an identity recognition system as described in any one of claims 1 to 6, as well as a main body and an irradiation unit; the main body is worn on the neck of the user; the hand image acquisition unit is arranged on the outer surface of the main body away from the neck; the irradiation unit is embedded in the surface of the main body close to the neck, and is used for laser irradiation of different positions of the neck; the image processing unit is arranged inside the main body, and the image processing unit is also communicated with the irradiation unit to drive the irradiation unit to enter a working state.

Images