Method for brushing equipment, brushing equipment and computer readable storage medium

Abstract

The invention discloses a method for a brushing device, the brushing device and a computer readable storage medium. The method comprises the steps that a palm image of a to-be-recognized target is acquired, the palm image at least comprises a mixed image, and the mixed image is a mixed image containing palm prints and palm veins at the same time; performing living body detection processing at least based on the mixed image to obtain a living body detection result of the palm image; and determining whether the to-be-identified target is a living body or not according to the living body detection result. Through the technical scheme disclosed by the invention, the living body detection can be realized at least based on the mixed image containing the palm print and the palm vein at the same time, so that the palm brushing equipment is prevented from being effectively attacked by some non-living bodies such as paper pictures, and the anti-attack capability and the security of the palm brushing equipment are improved.

Description

Technical Field

[0001] The present disclosure generally relates to the field of image processing technology. More specifically, the present disclosure relates to a method for a palm-swiping device, and a palm-swiping device and a computer-readable storage medium for executing the aforementioned method. Background Art

[0002] With the development of technology, biometric technology has been widely used in many fields such as identity authentication, gates, access control, attendance, etc. Among them, palm print and palm vein recognition technology has won the favor of users with its relatively high accuracy and stability. However, traditional biometric devices usually need to integrate RGB cameras for palm print recognition and IR cameras (i.e. infrared cameras) for palm vein recognition in the device, which not only makes the device have certain challenges in terms of size, cost, energy consumption, etc. More importantly, this method based on palm print or palm vein recognition has the hidden dangers of insufficient defense capabilities and easy to be attacked by non-living objects, making the security of the entire device insufficient. Summary of the invention

[0003] In order to at least solve one or more of the technical problems mentioned above, the present disclosure proposes solutions in multiple aspects that can improve the anti-attack capability of the palm-swiping device.

[0004] In a first aspect of an embodiment of the present disclosure, a method for a palm swiping device is provided, comprising: acquiring a palm image of a target to be identified, wherein the palm image comprises at least a mixed image, and the mixed image is a mixed image that contains both palm prints and palm veins; performing liveness detection processing based at least on the mixed image to obtain a liveness detection result of the palm image; and determining whether the target to be identified is alive based on the liveness detection result.

[0005] In some embodiments, the palm-swiping device is configured with a camera and an image enhancement component, wherein the image enhancement component is arranged relative to the position of the camera and is configured to adjust the exposure of the camera when capturing an image so that the palm image is captured, wherein the palm image includes one or more of the mixed image, palm print image, and palm vein image.

[0006] In some embodiments, performing liveness detection processing based at least on the mixed image includes: in response to the palm image only containing the mixed image, determining an area of ​​interest image about the mixed image, and performing liveness detection processing based on the area of ​​interest image of the mixed image; or in response to the palm image containing multiple images, determining an area of ​​interest image about each image, and performing liveness detection processing based on the area of ​​interest image of each image.

[0007] In some embodiments, performing liveness detection processing on the image of the region of interest of the mixed image includes: performing liveness detection processing on the image of the region of interest of the mixed image based on a pre-trained liveness recognition model.

[0008] In some embodiments, determining the area of ​​interest image for each image includes: performing image processing on each of the images to obtain the area of ​​interest image for each of the images; or performing image processing on any one of the multiple images to obtain the area of ​​interest image for any one of the images; and for other images among the multiple images, multiplexing the area of ​​interest in the area of ​​interest image for any one of the images into the other images to obtain the area of ​​interest images for the other images.

[0009] In some embodiments, performing liveness detection processing based on the region of interest image of each image includes: performing image fusion processing on the region of interest images of the multiple images to obtain a fused image; and performing liveness detection processing on the fused image based on a pre-trained liveness recognition model.

[0010] In some embodiments, performing image fusion processing on the region-of-interest images of the multiple images includes: performing channel stacking processing on the region-of-interest images of each of the images to obtain the fused image.

[0011] In some embodiments, performing liveness detection processing based on the region of interest image of each image includes: inputting the region of interest image of each image into the corresponding pre-trained liveness recognition model for processing to obtain the detection result of each image; and combining the detection results of each image to determine the liveness detection result of the palm image.

[0012] In some embodiments, determining the liveness detection result of the palm image includes: in response to the detection result of any image being non-liveness, determining the liveness detection result of the palm image is non-liveness; or in response to the detection results of all images being alive, determining the liveness detection result of the palm image is alive.

[0013] In some embodiments, the method further includes: detecting whether the image quality of the palm image meets a predetermined condition; and in response to determining that the image quality of the palm image meets the preset condition, executing a step of performing liveness detection processing based on at least the mixed image; or in response to determining that the image quality of the palm image does not meet the preset condition, adjusting the exposure of the palm prints and / or palm veins in the image captured by the camera through the image enhancement component to recapture the palm image until the image quality of the new palm image meets the preset condition.

[0014] In some embodiments, the method further includes: in response to determining that the target to be identified is a living body, performing identity recognition processing based on the palm image.

[0015] In a second aspect of the disclosed embodiment, a palm-brushing device is provided, further comprising: a processor; and a memory storing computer instructions for the palm-brushing device, wherein when the computer instructions are executed by the processor, the palm-brushing device executes the method described in the foregoing and following embodiments.

[0016] In a third aspect of the disclosed embodiments, a computer-readable storage medium is provided, comprising program instructions for a palm-swiping device, and when the program instructions are executed by a processor, the method described in the foregoing and following embodiments is implemented.

[0017] Through the solution for palm scanning device provided above, the disclosed embodiment can obtain a mixed image of the target to be identified, and detect whether the target to be identified is a living body based on at least the mixed image containing both palm prints and palm veins. It can be seen that the disclosed solution realizes liveness detection through a mixed image that is difficult to imitate, avoids the palm scanning device from being effectively attacked by some non-living objects such as paper pictures, and improves the anti-attack capability and security of the palm scanning device. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By reading the detailed description below with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood. In the accompanying drawings, several embodiments of the present disclosure are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0019] Figure 1 A schematic flow chart of a method for a palm-brushing device according to an embodiment of the present disclosure is shown;

[0020] Figure 2 A structural block diagram of a palm brushing device according to an embodiment of the present disclosure is shown;

[0021] Figure 3 A schematic flow chart showing a method for a palm brushing device according to another embodiment of the present disclosure is shown;

[0022] Figure 4 A schematic flow chart of a method for a palm-brushing device according to yet another embodiment of the present disclosure is shown;

[0023] Figure 5 A schematic flow chart of a method for performing liveness detection based on palm prints and / or palm veins according to an embodiment of the present disclosure is shown;

[0024] Figure 6A schematic flow chart showing a method for performing liveness detection based on palm prints and / or palm veins according to another embodiment of the present disclosure; and

[0025] Figure 7 A schematic structural diagram of a palm-brushing device according to another embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0027] It should be understood that the terms "include" and "comprising" used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0028] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms of "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations.

[0029] As used in this specification and claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0030] Generally, devices that support palm print and palm vein recognition usually need to integrate RGB cameras and IR cameras in the devices, and collect palm print images and palm vein images based on the RGB cameras and IR cameras respectively, and need to perform palm print and palm vein recognition based on the palm print images and palm vein images respectively. In this regard, the prior art also has a design that only sets one camera in the device, and this device can collect palm print images and palm vein images in different time periods through one camera. Although this design reduces the size of the device to a certain extent, in essence, it still needs to collect palm print images and palm vein images separately, and perform palm print and palm vein recognition based on the palm print images and palm vein images.

[0031] However, in actual applications, such palm print images or palm vein images are easily forged (for example, printed on drawings, etc.), making the device vulnerable to attacks by non-living objects (such as drawings), and the anti-attack capability is poor, posing certain security risks.

[0032] In this regard, the inventor has found through research that a mixed image that can contain both palm prints and palm veins is difficult to copy or forge. By implementing liveness detection through such a mixed image that is difficult to forge, it is possible to prevent the palm scanning device from being effectively attacked by some non-living objects such as paper pictures.

[0033] The specific implementation of the present disclosure is described in detail below with reference to the accompanying drawings.

[0034] Figure 1 A schematic flow chart of a method 100 for a palm scanning device according to an embodiment of the present disclosure is shown. It should be noted that the palm scanning device in the present disclosure may include any device that has palm print and / or palm vein recognition requirements and has the function of collecting a mixed image containing both palm prints and palm veins.

[0035] like Figure 1 As shown, at step S101, a palm image of the target to be identified can be obtained. The palm image at least includes a mixed image, which is an image containing both palm prints and palm veins. Usually, the palm, palm prints and palm veins are unique, and the palm prints are located on the surface of the palm, and the palm veins are located under the skin of the palm. The relative position or superposition of the two in the palm is unique. Even if the palm print image and the palm vein image can be obtained, it is difficult to use the random combination of the palm print image and the palm vein image to imitate an effective mixed image. Therefore, this example needs to collect at least a mixed image of the target to be identified.

[0036] In practical applications, the palm image of the target to be identified can be obtained in a variety of ways. For example, when the target to be identified is detected to appear in the image acquisition area of ​​the palm-brushing device, the palm-brushing device can be automatically triggered to at least collect a mixed image of the target to be identified. For another example, when the palm-brushing device receives a user instruction (such as a voice instruction or other control instruction, etc.), it collects a mixed image of the target to be identified. It should be noted that the description of the palm image acquisition process here is only an exemplary description, and can be adjusted according to the requirements of human-computer interaction design.

[0037] At step S102, liveness detection processing may be performed based on at least the mixed image to obtain a liveness detection result of the palm image. In some embodiments, when only the mixed image is collected, liveness detection processing may be performed based on the mixed image. In other instances, in addition to the mixed image, other images (such as palm print images and / or palm vein images of the target to be identified) may also be collected. In this case, liveness detection processing may be performed based on the mixed image and other images. In this way, liveness detection requirements of different application scenarios can be met.

[0038] At step S103, it can be determined whether the target to be identified is a living body based on the above liveness detection result. In some embodiments, when liveness detection is performed only on the mixed image, it can be determined whether the target to be identified is a living body based on the liveness detection result of the mixed image. In other embodiments, when liveness detection is performed on the mixed image and other images, it is necessary to determine whether the target to be identified is a living body based on the liveness detection results of the mixed image and other images.

[0039] Therefore, liveness detection is achieved through a mixed image that is difficult to imitate, which prevents the palm scanning device from being effectively attacked by non-living objects such as paper pictures, thereby improving the anti-attack capability and security of the palm scanning device.

[0040] The above method 100 is applicable to a palm brush device. As mentioned above, the palm brush device is a device that has the function of collecting mixed images. In the specific technical implementation process, the palm brush device may collect mixed images in a variety of ways. However, this disclosure defines a new type of palm brush device, Figure 2 A possible structural block diagram of the palm-brushing device 200 according to an embodiment of the present disclosure is shown.

[0041] like Figure 2 As shown, the palm scanning device 200 is configured with a camera 201 and an image enhancement component 202. The image enhancement component 202 is arranged relative to the position of the camera 201 and is configured to adjust the exposure of the camera 201 when collecting images so as to collect palm images. The palm image may include one or more images of a mixed image, a palm print image, and a palm vein image.

[0042] Considering the different positions of palm prints and palm veins in the palm, and the different exposure requirements of the camera when collecting palm prints and palm veins, it is generally necessary to collect palm print images and palm vein images separately. After further research, the inventor also found that in actual applications, although RGB images of palm prints are generally collected, palm print information is mainly presented through grayscale images, that is, grayscale images of palm prints can meet the use requirements, and palm vein information is also mainly presented through grayscale images. Therefore, a single camera that can support grayscale image acquisition can be used, and by adjusting the exposure of the camera when collecting images, it can be used to collect a mixed grayscale image (hereinafter referred to as a mixed image) to simultaneously present palm prints and palm veins based on the mixed image. As a result, there is no need for multiple cameras, nor is there a need to collect palm print images and palm vein images separately, which effectively reduces the size of the device.

[0043] The camera 201 and the image enhancement component 202 in the present disclosure constitute an image processing module. In an image processing module, the camera 201 is a single camera and can be configured to collect grayscale images of palm prints or palm veins. It should be noted that the present disclosure does not limit the specific type of camera, so any camera with the function of collecting grayscale images can be used.

[0044] As mentioned above, the location and exposure requirements of palm prints and palm veins are different. Among them, the palm prints are located on the surface of the palm, and images containing the palm prints can be collected when exposed in a visible light environment. The palm veins are located under the skin of the palm, and images containing the palm veins can be collected when exposed in an infrared light environment. Therefore, this embodiment adds an image enhancement component 202, and uses the image enhancement component 202 to adjust the exposure of the camera 201 when collecting images, so that the proportion of infrared light and / or visible light entering the camera 201 is moderate, so that the palm prints and / or palm veins can be presented in different degrees of exposure in the image. In particular, when an appropriate proportion of infrared light and visible light enter the camera, the camera can collect a mixed image that contains both palm prints and palm veins.

[0045] This embodiment integrates a single camera in the image processing module of the palm scanning device, so that the overall volume is effectively reduced, and the palm print and palm vein can be simultaneously presented through mixed images. Therefore, it can be seen that the solution disclosed in this disclosure not only breaks through the traditional thinking inertia of collecting palm print and palm vein images separately, but also develops a new image processing idea while meeting the use requirements.

[0046] Furthermore, in some embodiments, the image enhancement component 201 may be specifically configured to adjust the exposure of the palm print and / or palm vein in the image, so that the camera 201 can capture a mixed image, a palm print grayscale image (hereinafter referred to as the palm print image) and / or a palm vein grayscale image (hereinafter referred to as the palm vein image). Through the cooperation between the camera 201 and the image enhancement component 202 in this embodiment, only the palm print can be exposed in the image to capture the palm print image. Only the palm vein can also be exposed in the image to capture the palm vein image, and only the palm print and the palm vein can also be exposed in the image at the same time to capture the mixed image, so as to meet different image acquisition requirements.

[0047] Further, in some embodiments, the image enhancement component 202 may include a shading sheet. The shading sheet is disposed in front of the camera 201 and is configured to adjust the light transmittance ratio of the infrared light and the visible light beam incident on the camera 201. The shading sheet here may be a shading sheet that transmits both infrared light and visible light. There is no restriction on the specific type of filter here. The shading sheet may be a filter with a fixed light transmittance ratio or an adjustable filter, and may be specifically set according to application requirements. For example, when the light in the application environment is stable and the image acquisition requirements are fixed (for example, only a certain type of image is acquired), a filter with a fixed light transmittance ratio may be used.

[0048] When the application environment and image acquisition requirements are changeable, an adjustable filter can be used. For example, electrochromic glass with dual pass for visible light and infrared light or other adjustable shading sheets can be used. In this scenario, in addition to the aforementioned shading sheet, the image enhancement component 202 also includes a first adjustment unit. The first adjustment unit is configured to obtain quality information of the image captured by the camera 201, and can adjust the light transmittance ratio of the infrared light and the visible light of the adjustable filter according to the quality information. In some embodiments, when an image (such as a palm print image, a palm vein image, or a mixed image) is captured, the image quality can be evaluated to obtain quality information, and the light transmittance ratio of the infrared light and the visible light of the adjustable filter can be adjusted based on the quality information.

[0049] The specific quality evaluation method and the specific process of adjusting the transmittance ratio of infrared light and visible light of the adjustable filter are not limited here. In practical applications, image quality evaluation can be achieved in a variety of ways. For example, a quality evaluation score threshold can be set, and meeting the threshold means that the image quality requirements (such as recognition and / or biopsy image quality requirements, etc.) are met. Among them, the threshold can be set according to the image clarity, and can also be set according to any other reasonable evaluation parameters. The specific adjustment method of the transmittance ratio can be obtained through deep learning, or it can be obtained by any reasonable method such as table lookup, fuzzy control, etc.

[0050] In other embodiments, the image enhancement component 202 may include a fill light group and a second adjustment unit.

[0051] Among them, the fill light group can be arranged relative to the position of the camera 201. The number of lights in the fill light group and the specific arrangement are not limited here, and can be set specifically according to application requirements. In addition, the lights in the fill light group can be evenly distributed around the camera in any arrangement. In some embodiments, the fill light group may include infrared fill light and visible light fill light, and the infrared fill light and visible light fill light can be evenly distributed around the camera in any way. For example, a circle of visible light fill light and a circle of infrared light fill light are arranged in sequence from the camera to the outside, or only one circle of fill light is set, in which the visible light fill light and the infrared light fill light are arranged at intervals, etc. Of course, if the visible light or infrared light in the application environment is sufficient, the fill light group may also only include infrared light fill light or visible light fill light.

[0052] The aforementioned second adjustment unit can be specifically configured to obtain quality information of the image captured by the camera, and adjust the current of the fill light in the fill light group according to the quality information. In some embodiments, when images such as palm print images, palm vein images, or mixed images are captured, the image quality can be evaluated to obtain quality information, and the current of the fill light in the fill light group can be adjusted based on the quality information.

[0053] In practical applications, image quality evaluation can be achieved in a variety of ways. For example, a quality evaluation score threshold can be set, and meeting the threshold means meeting the image quality requirements (such as recognition and / or biopsy image quality requirements, etc.). The threshold can be set according to the image clarity, or according to any other reasonable evaluation parameters. The specific adjustment method for the current of the fill light can be obtained through deep learning, or it can be obtained by any reasonable method such as table lookup, fuzzy control, etc. The specific quality evaluation method and the specific process of how to adjust the current of the fill light in the fill light group are not limited here.

[0054] In some embodiments, the aforementioned fill light group may include an infrared fill light and a visible fill light, and the second adjustment unit is specifically configured to adjust the current ratio between the infrared fill light and the visible fill light according to the quality information. For example, if it is detected that the palm vein clarity in the image is insufficient, the current ratio between the infrared fill light and the visible fill light can be increased so that the palm can be fully exposed to infrared light, thereby improving the palm vein clarity. When it is detected that the palm print clarity in the image is insufficient, the current ratio between the infrared fill light and the visible fill light can be reduced so that the palm can be fully exposed to visible light, thereby improving the palm print clarity.

[0055] In some other embodiments, the image enhancement component 202 may specifically include a shading sheet, a fill light group, a first adjustment unit, and a second adjustment unit. In practical applications, the exposure of the camera 201 when acquiring an image may be adjusted based on the shading sheet, the first adjustment unit, the fill light group, and the second adjustment unit in the image enhancement component 202. The exposure of the camera 201 when acquiring an image may also be adjusted based on the shading sheet and the first adjustment unit in the image enhancement component. The exposure of the camera when acquiring an image may also be adjusted based on the fill light group and the second adjustment unit in the image enhancement component 202. For details about the shading sheet, the fill light group, the first adjustment unit, and the second adjustment unit, please refer to the previous description, which will not be repeated here.

[0056] Thus, through the cooperation between the image enhancement component and the single camera in the palm scanning device, any one or more images among the mixed image, palm print image, and palm vein image can be collected to meet the requirements of different application scenarios. In addition, the setting of the first adjustment unit and / or the second adjustment unit in the image enhancement component can enable the image processing module to adapt the proportion of visible light and infrared light in different application scenarios according to the requirements of different scenarios, so as to obtain a clear palm print image, palm vein image, or a palm print and palm vein mixed image, thereby increasing the scene adaptation capability of the image processing module and improving its recognition accuracy in various scenarios.

[0057] It should be noted that the number of image processing modules that can be integrated into the palm brush device 200 is not limited here, and can be adjusted according to application requirements. For example, for some palm brush devices that require a compact structure, only one image processing module can be integrated. For some palm brush devices that require powerful image functions, multiple image processing modules can be integrated to improve the image parallel processing capability of the entire palm brush device.

[0058] The following combination Figure 3 and Figure 4 The specific implementation process of the palm-brushing device method is described. Figure 3 and Figure 4 The methods involved can be applied to Figure 2 The palm scanning device 200 shown, or other devices capable of collecting mixed images, palm print images and palm vein images. In addition, Figure 3 and Figure 4 The methods involved can be understood as the specific implementation and further development of method 100. Figure 1 and Figure 2 The relevant description also applies to the following.

[0059] Figure 3FIG. 4 is a schematic flow chart of a method 300 for palm-swiping a device according to an embodiment of the present disclosure.

[0060] Specifically, Figure 3 As shown, at step S301, a mixed image of the target to be identified can be obtained. In some embodiments, if the fill light group of the palm scanning device is provided with a visible light fill light and an infrared light fill light, the visible light fill light and the infrared light fill light can be turned on at the same time to achieve joint exposure of the palm print and the palm vein to collect a mixed image. It should be noted that the description of the mixed image acquisition process here is only an exemplary description, and the disclosed solution is not limited to this.

[0061] At step S302, in response to acquiring the mixed image, a region of interest image about the mixed image can be determined. In some embodiments, palm detection and key point positioning can be performed on the mixed image, and after obtaining the corresponding detection frame and key points, the image can be further normalized by rotation, cropping, etc. to obtain a region of interest image (ROI image) of the mixed image. Among them, deep learning technology or other image positioning technology can be used to achieve palm detection and key point positioning, and the description of the region of interest image determination process here is only an exemplary description.

[0062] At step S303, it can be detected whether the image quality of the mixed image meets the predetermined conditions. If the predetermined conditions are met, step S305 is executed. If the predetermined conditions are not met, step S304 is executed. The evaluation method for image quality can include multiple methods. For example, it can be evaluated whether the image has image defects such as being too dark, overexposed, having areas blocked, and abnormal posture, and / or it can also be determined based on the image quality whether the ratio of infrared light to visible light needs to be adjusted. Among them, the ratio of infrared light to visible light can be adjusted by adjusting the ratio of the light transmittance of infrared light and visible light in the adjustable filter and / or adjusting the current ratio between the infrared fill light and the visible light fill light. The predetermined conditions here can be set and adjusted in combination with the specific evaluation method of image quality.

[0063] At step S304, the exposure of the palm prints and palm veins in the image captured by the camera can be adjusted by the image enhancement component in the palm brush device. In some embodiments, the application scenario requirements of the palm brush device can be determined based on obtaining the natural light intensity information of the environment in which the palm brush device is located and / or obtaining the working mode of the palm brush device. And in combination with the application scenario requirements, the ratio of the light transmittance of infrared light and visible light in the adjustable filter in the image enhancement component and / or the current ratio between the infrared fill light and the visible light fill light are adjusted to achieve the exposure adjustment of the palm prints and palm veins in the image. After the adjustment is completed, return to execute step S301 to re-capture the mixed image until the image meets the predetermined conditions.

[0064] At step S305, liveness detection processing can be performed on the image of the region of interest of the mixed image based on the pre-trained liveness recognition model. In practical applications, the liveness recognition model can be trained using deep learning technology. For example, the liveness recognition model can be obtained by training some basic image processing network models using sample images. It should be noted that the description of the liveness recognition model here is only an exemplary description.

[0065] In step S306, in response to determining that the target to be identified is a living body, identity recognition processing can be performed based on the mixed image. The image of the region of interest of the mixed image is input into the liveness recognition model for processing, and the liveness recognition result can be output. If the liveness recognition result is a live body, it is determined that the target to be identified is a live body, and the mixed image can be used for subsequent identity recognition processing. If the liveness recognition result is not a live body, the subsequent identity recognition processing is not performed to avoid wasting computing resources. At the same time, alarm prompts can also be given by voice, sound and light. In this way, liveness detection based on mixed images is realized, and the anti-attack capability and security of the palm-swiping device are improved.

[0066] Figure 4 FIG. 4 is a schematic flow chart of a method 400 for a palm-swiping device according to another embodiment of the present disclosure. It should be noted that the method 300 exemplarily demonstrates the process of performing liveness detection based on a mixed image, while the method 400 demonstrates the process of performing liveness detection based on a mixed image and other images. Figure 3 The detailed description about mixed images in also applies below.

[0067] like Figure 4 As shown, in step S401, multiple images of the mixed image, palm print image and palm vein image of the target to be identified can be obtained. Considering that the palm-swiping device may need to be applied to application scenarios with higher security requirements, multiple images can be collected based on the palm-swiping device at this time, so that liveness detection can be performed based on multiple images later, further improving the anti-attack capability of the palm-swiping device.

[0068] The multiple images in this embodiment may include a mixed image and a palm print image, a mixed image and a palm vein image, a mixed image, a palm print image and a palm vein image, etc.

[0069] In some embodiments, if a visible light fill light and an infrared light fill light are provided in the fill light group of the palm-swiping device, the visible light fill light and the infrared light fill light can be turned on at the same time to realize the joint exposure of the palm print and the palm vein to collect a mixed image. The visible light fill light can be turned on and the infrared light fill light can be turned off to realize the collection of the palm print image. The visible light fill light can be turned off and the infrared light fill light can be turned on to realize the collection of the palm vein image. It should be noted that the description of the acquisition process of each image here is only an exemplary description, and the scheme disclosed herein is not limited to this. For example, this embodiment does not limit the specific acquisition order, quantity, etc. of each image.

[0070] At step S402, an area of ​​interest image for each image may be determined. In some embodiments, each image may be processed to obtain an area of ​​interest image for each image. In other embodiments, when the palm-brushing device uses a single camera for image acquisition, and the exposure time is short and there is no obvious change in the target to be identified, image processing may be performed on only any one of the aforementioned multiple images to obtain an area of ​​interest image for any image. And for other images in the multiple images, the area of ​​interest in the area of ​​interest image of any image is reused in other images to obtain area of ​​interest images for other images. In this way, the entire image processing process can be simplified and computing resources can be saved. Among them, the process of determining the area of ​​interest image of each image can refer to Figure 3 The relevant description in will not be repeated here.

[0071] At step S403, it can be detected whether the image quality of the acquired image meets the predetermined condition. If the predetermined condition is met, step 405 is executed. If the predetermined condition is not met, step S404 is executed. The acquired image here can include each of the aforementioned multiple images, or any image of the multiple images. The image quality evaluation method for each image or any image can include multiple methods. The specific evaluation method can be referred to in the previous text. Figure 3 The relevant description in will not be repeated here.

[0072] At step S404, the exposure of the palm prints and / or palm veins in the image captured by the camera can be adjusted by the image enhancement component in the palm scanning device, and the image that does not meet the predetermined condition can be recaptured until the image meets the predetermined condition. The exposure adjustment process of the palm prints and / or palm veins in the image can be referred to in the previous Figure 3 The relevant description in will not be repeated here.

[0073] In step S405, living body detection processing may be performed based on the region of interest image of each type of image.

[0074] In some embodiments, the image fusion processing can be performed on the region of interest images of the aforementioned multiple images to obtain a fused image. For example, the channel stacking processing can be performed on the region of interest images of each image to obtain a multi-channel fused image. Then, the fused image is subjected to liveness detection processing based on a pre-trained liveness recognition model. Among them, the liveness recognition model can be obtained by training some basic image processing models. Then, according to the liveness detection result of the liveness recognition model on the fused image, it is determined whether the target to be identified is a live body.

[0075] In other embodiments, the region of interest image of each image can be input into the corresponding pre-trained liveness recognition model for processing to obtain the detection result of each image. Each activity recognition model here can be obtained by training some basic image processing models. Then, the detection result of each image is combined to determine the liveness detection result of the palm image. Specifically, in response to the detection result of any image being non-live, the liveness detection result of the palm image is determined to be non-live. In response to the detection results of all images being live, the liveness detection result of the palm image is determined to be live.

[0076] In step S406, in response to determining that the target to be identified is a living body, identity recognition processing can be performed based on the acquired image. Thus, living body detection based on multiple images is achieved to meet the detection requirements of different application scenarios.

[0077] Figure 5 FIG. 5 is a schematic flow chart of a method 500 for performing liveness detection based on palm prints and / or palm veins according to an embodiment of the present disclosure. It can be understood that method 500 is a further limitation or extension of method 100 and method 300. Figure 1 and Figure 3 The relevant detailed description also applies to the following.

[0078] like Figure 5 As shown, at step S501, the distance between the target to be identified in front of the lens and the lens can be detected, and the distance information can be output. In this example, a single camera and an image enhancement component including a shading sheet and a fill light group can be set on the palm brush device. Among them, the shading sheet can be a visible light infrared dual-pass filter, and the fill light group can include an infrared fill light and a visible light fill light. Furthermore, a distance sensor can also be set on the palm brush device. For example, an infrared distance sensor, etc. Based on the infrared distance sensor, the distance between the object in front of the lens and the lens can be detected, and the distance information can be output.

[0079] In step S502, after detecting that the target to be identified falls within a preset distance (the preset distance can be set and adjusted according to needs), the infrared light fill light and the visible light fill light are turned on at the same time to achieve joint exposure to collect a mixed image.

[0080] At step S503, palm detection and key point positioning may be performed based on the mixed image. For example, deep learning technology or other image positioning technology may be used to implement palm detection and key point positioning.

[0081] At step S504, the region of interest image of the mixed image may be obtained. For example, after performing palm detection and key point positioning on the mixed image and obtaining the corresponding detection frame and key points, the image may be further normalized by rotation, cropping, etc. to obtain the region of interest image of the mixed image.

[0082] At step S505, image quality judgment may be performed based on the region of interest of the mixed image.

[0083] For example, it is possible to evaluate whether the image has image defects such as being too dark, overexposed, having areas blocked, and abnormal posture, and / or it is also possible to determine whether the ratio of infrared light to visible light needs to be adjusted based on the image quality to achieve image quality judgment. When it is determined that the image quality does not meet the standard, it is necessary to adjust the ratio of infrared light to visible light until the image quality meets the standard. The specific adjustment process of the ratio of infrared light to visible light can refer to the description of the relevant details in the previous text, and will not be repeated here.

[0084] At step S506, after determining that the image quality meets the standard, liveness detection can be performed based on the image of the region of interest of the mixed image. Specifically, liveness detection processing can be performed on the image of the region of interest of the mixed image based on a pre-trained liveness recognition model. For example, the sample images can be used to train some basic image processing network models (such as models of the mobilenet series, shufflenet series, etc.) to obtain the liveness recognition model. It should be noted that the description of the liveness recognition model here is only an exemplary description.

[0085] In step S507, in response to determining that the target to be identified is a living body, palm print and palm vein identification can be performed. Specifically, the image of the region of interest of the mixed image can be input into the living body identification model for processing, and the living body identification result can be output. If the living body identification result is a living body, it is determined that the target to be identified is a living body, and the mixed image can be used for subsequent palm print and palm vein identification processing.

[0086] Figure 6FIG. 6 is a schematic flow chart of a method 600 for performing liveness detection based on palm prints and / or palm veins according to an embodiment of the present disclosure. It can be understood that method 600 is a further limitation or extension of method 100 and method 400. Figure 1 and Figure 4 The relevant detailed description also applies to the following.

[0087] like Figure 6 As shown, at step S601, the distance between the target to be identified in front of the lens and the lens can be detected, and the distance information can be output. In this example, a single camera and an image enhancement component including a shading sheet and a fill light group can be set on the palm brush device. Among them, the shading sheet can be a visible light infrared dual-pass filter, and the fill light group can include an infrared fill light and a visible light fill light. Furthermore, a distance sensor can also be set on the palm brush device. For example, an infrared distance sensor, etc. Based on the infrared distance sensor, the distance between the object in front of the lens and the lens can be detected, and the distance information can be output.

[0088] At step S602, after determining that the target is at the preset distance, the palm print image, the palm vein image and the mixed image are collected respectively. Specifically, after detecting that the target to be identified falls within the preset distance, the visible light fill light is turned on and the infrared fill light is turned off, and the palm print image is collected based on the camera, the visible light fill light is turned off and the infrared fill light is turned on, and the palm vein image is collected based on the camera, and the visible light fill light and the infrared fill light are turned on and the mixed image is collected based on the camera. The description of the collection process of each image here is only an exemplary description. In actual application, the order of obtaining the above three images is not limited.

[0089] At step S603, palm detection and key point location can be performed based on the palm print image, palm vein image and mixed image. Specifically, palm detection and key point location can be performed on the above three images respectively. Or when the palm scanning device uses a single camera for image acquisition, the exposure time is short and there is no obvious movement of the palm, palm detection and key point location can be performed on only any one of the above three images, and the other images can reuse the detection frame and key points of any one of the images for subsequent determination of the area of ​​interest image.

[0090] At step S604, the region of interest images (ROI images) of the palm print image, the palm vein image, and the mixed image may be obtained. For example, after the detection frames and key points of the palm print image, the palm vein image, and the mixed image may be obtained, the images may be further normalized by rotation, cropping, etc. to obtain the region of interest images corresponding to the palm print image, the palm vein image, and the mixed image, respectively.

[0091] At step S605, the image quality can be judged based on the image of the region of interest of any image. Specifically, it can be detected whether the image quality of the image meets the predetermined conditions. If the predetermined conditions are not met, the exposure of the palm prints and / or palm veins in the image captured by the camera can be adjusted by the image enhancement component in the palm brushing device, and the image that does not meet the predetermined conditions can be recaptured until the image meets the predetermined conditions. Among them, the specific evaluation process of the image quality and the exposure adjustment process of the palm prints and palm veins in the image can refer to the relevant description above, which will not be repeated here.

[0092] In step S606, after determining that the image quality meets the standard, liveness determination can be performed based on the palm print image, the palm vein image, and the mixed image.

[0093] Specifically, after obtaining the palm vein ROI image, palm print ROI image and mixed ROI image, two modes, image-level fusion and result-level fusion, can be used to integrate the above images into the liveness detection algorithm to enhance the anti-attack capability of the recognition system. For image-level fusion, the palm vein ROI image, palm print ROI image and mixed ROI image can be stacked by channels before inputting into the liveness recognition model. For example, the three images can all be regarded as single-channel images, and the three single-channel images can be stacked by channels to obtain a three-channel image. The three-channel image is then used as the input of the liveness recognition model, and the result obtained by the liveness recognition model is the result of the entire liveness detection.

[0094] For the fusion of the result level, the palm vein ROI image, the palm print ROI image and the mixed ROI image can be used as the input of the liveness recognition model A, the liveness recognition model B and the liveness recognition model C respectively. Then, the result of the entire liveness detection can be obtained by combining the recognition results of the liveness recognition model A, the liveness recognition model B and the liveness recognition model C. For example, if the liveness recognition model A, the liveness recognition model B and the liveness recognition model C are all judged as live, then the result of the entire liveness detection is determined to be live. In other cases, the result of the liveness detection is determined to be a fake (that is, non-live).

[0095] In step S607, after determining that the target to be identified is a living body, subsequent palm print and palm vein identification may be performed using the above image.

[0096] After introducing the method of the exemplary embodiment of the present disclosure, next, reference is made to Figure 7 A palm brushing device according to an exemplary embodiment of the present disclosure is described.

[0097] Figure 7 A schematic block diagram of a palm-brushing device 700 according to another embodiment of the present disclosure is schematically shown.

[0098] like Figure 7 As shown, the palm brushing device 700 may include a processor 701 and a memory 702. The memory 702 stores computer instructions for the palm brushing device. When the computer instructions are executed by the processor 701, the palm brushing device 700 executes the above-mentioned Figure 1 , Figure 3 to Figure 6 It should be noted that the palm brushing device 700 in this embodiment may have Figure 2 The structure of the palm scanning device 200 may also include other devices capable of collecting mixed images, palm print images and palm vein images. The palm scanning device 700 can implement liveness detection based at least on the mixed image that is difficult to forge, thereby preventing the palm scanning device from being effectively attacked by some non-living objects such as paper pictures, thereby improving the anti-attack capability and security of the palm scanning device.

[0099] In addition, the present disclosure also provides a computer-readable storage medium in which program instructions are stored, and the program instructions are configured to be executed at runtime. Figure 1 , Figure 3 to Figure 6 A method for a palm-swiping device as shown in any of the figures.

[0100] Specifically, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0101] It should be noted that although several devices or sub-devices of the device are mentioned in the above detailed description, this division is only exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more devices described above can be embodied in one device. Conversely, the features and functions of one device described above can be further divided into multiple devices to be embodied.

[0102] Although the spirit and principles of the present disclosure have been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the disclosed specific embodiments, and the division of various aspects does not mean that the features in these aspects cannot be combined to benefit, and such division is only for the convenience of expression. The present disclosure is intended to cover various modifications and equivalent arrangements included in the spirit and scope of the attached claims. The scope of the attached claims conforms to the broadest interpretation, thereby including all such modifications and equivalent structures and functions.

Claims

1. A method for a palm brushing device, characterized in that: include: Acquire a palm image of a target to be identified, wherein the palm image at least includes a mixed image, and the mixed image is a mixed image that includes both palm prints and palm veins; Performing liveness detection processing based at least on the mixed image to obtain a liveness detection result of the palm image; and According to the liveness detection result, it is determined whether the target to be identified is a live body.

2. The method according to claim 1, characterized in that The palm-swiping device is provided with a camera and an image enhancement component, wherein the image enhancement component is arranged relative to the position of the camera and is configured to adjust the exposure of the camera when capturing an image so as to capture the palm image, wherein the palm image includes one or more of the mixed image, palm print image and palm vein image.

3. The method according to claim 2, characterized in that Performing liveness detection processing based at least on the mixed image includes: In response to the palm image only including the mixed image, determining an area-of-interest image about the mixed image, and performing living body detection processing based on the area-of-interest image of the mixed image; or In response to the palm image including a plurality of images, an interest region image for each image is determined, and a living body detection process is performed based on the interest region image for each image.

4. The method according to claim 3, characterized in that: Performing a living body detection process based on the region of interest image of the mixed image includes: Based on the pre-trained living body recognition model, living body detection processing is performed on the region of interest image of the mixed image.

5. The method according to claim 3, characterized in that: Determining the region of interest for each image includes: Performing image processing on each of the images to obtain an image of a region of interest for each of the images; or Performing image processing on any one of the multiple images to obtain an image of a region of interest of the any one of the images; and For other images among the multiple images, the region of interest in the region of interest image of any one of the images is multiplexed into the other images to obtain the region of interest images related to the other images.

6. The method according to claim 3, characterized in that Performing a living body detection process based on the region of interest image of each of the images includes: performing image fusion processing on the interest region images of the multiple images to obtain a fused image; and The fused image is subjected to liveness detection processing based on a pre-trained liveness recognition model.

7. The method according to claim 6, characterized in that Performing image fusion processing on the interest region images of the multiple images includes: Channel stacking processing is performed on the region of interest image of each of the images to obtain the fused image.

8. The method according to claim 3, characterized in that Performing a living body detection process based on the region of interest image of each of the images includes: Inputting the region of interest image of each type of image into the corresponding pre-trained living body recognition model for processing to obtain the detection result of each type of image; and The detection results of each type of image are combined to determine the liveness detection result of the palm image.

9. The method according to claim 3, characterized in that: Determining a liveness detection result of the palm image includes: In response to a detection result of any one of the images being non-live, determining a liveness detection result of the palm image as non-liveness; or In response to the detection results of all the images being alive, it is determined that the liveness detection result of the palm image is liveness.

10. The method according to claim 3, characterized in that The method further comprises: detecting whether the image quality of the palm image meets a predetermined condition; and In response to determining that the image quality of the palm image meets the preset condition, executing a step of performing a living body detection process based at least on the mixed image; or In response to determining that the image quality of the palm image does not meet the preset condition, the exposure of the palm prints and / or palm veins in the image captured by the camera is adjusted by the image enhancement component to recapture the palm image until the image quality of the new palm image meets the preset condition.

11. The method according to any one of claims 1 to 10, characterized in that: The method further comprises: In response to determining that the target to be identified is a living body, identity recognition processing is performed based on the palm image.

12. A palm brushing device, characterized in that: Also includes: processor; as well as A memory storing computer instructions for a palm-brushing device, wherein when the computer instructions are executed by the processor, the palm-brushing device executes the method according to any one of claims 1 to 11.

13. A computer-readable storage medium, characterized in that: The invention comprises program instructions for a palm-swiping device, and when the program instructions are executed by a processor, the method according to any one of claims 1 to 11 is implemented.

Images