Infrared palm living body detection method, system and device and storage medium

Abstract

The invention discloses an infrared palm living body detection method, system and device and a storage medium. The method comprises the steps that S1, an infrared palm image is acquired; s2, performing Fourier transform on the infrared palm image to obtain frequency spectrum information; s3, extracting a high-frequency part and a low-frequency part in the frequency spectrum information; s4, performing Fourier inverse transformation on the high-frequency part to obtain a high-frequency infrared palm image; inverse Fourier transform is carried out on the low-frequency part to obtain a low-frequency infrared palm image; and S5, carrying out living body identification on the high-frequency infrared palm image and the high-frequency infrared palm image by using a living body identification model. According to the invention, the palm can be accurately subjected to living body recognition.

Description

Background Art

[0002] With the wide application of biometric technologies, palm recognition, as an efficient means of identity authentication, has gradually attracted attention. However, palm recognition systems face the risk of spoofing attacks, such as being deceived by means of photos, videos, or molds. Therefore, liveness detection technology has become a key link to ensure the security of palm recognition systems.

[0003] Infrared imaging technology utilizes the thermal radiation characteristics of objects and can capture the physiological characteristics of the human body, such as blood flow and temperature distribution. Compared with visible light imaging, infrared images are less dependent on environmental lighting and can effectively distinguish between live and non-live objects. For example, the temperature distribution and blood flow of a live palm will produce unique infrared features, while a forged palm cannot reproduce these features.

[0004] Fourier transform is a mathematical tool that transforms an image from the spatial domain to the frequency domain and can extract the frequency characteristics of the image. In liveness detection, the high-frequency part usually reflects the details and textures of the image, while the low-frequency part reflects the overall structure and temperature distribution. By extracting these features through Fourier transform, the palm image can be analyzed more comprehensively, thereby improving the accuracy of liveness detection.

[0005] In recent years, deep learning technology has made remarkable progress in the fields of image recognition and liveness detection. By training a deep learning model, the key features in infrared images can be automatically learned and extracted, thus achieving efficient and accurate liveness detection. For example, a convolutional neural network (CNN) can process complex image data and classify live and non-live features.

[0006] In practical applications, palm liveness detection technology has a wide range of application prospects. For example, in access control systems, liveness detection can prevent unauthorized personnel from entering restricted areas by forging palms. In financial payment scenarios, liveness detection can ensure the security of transactions and prevent fraud. Therefore, developing a palm liveness detection method based on infrared imaging and Fourier transform is of great significance for enhancing the security of biometric systems.

[0007] The disclosure of the above background art content is only for assisting in understanding the inventive concept and technical solution of the present invention, and it does not necessarily belong to the prior art of this patent application. Without clear evidence indicating that the above content was publicly available on the filing date of this patent application, the above background art should not be used to evaluate the novelty and inventiveness of this application. Summary of the Invention

[0008] To this end, the present invention provides an infrared palm liveness detection solution to meet the need for detecting palm liveness in various application scenarios.

[0009] In a first aspect, the present invention provides an infrared palm liveness detection method, which is characterized by comprising:

[0010] Step S1: Obtain an infrared palm image;

[0011] Step S2: Perform a Fourier transform on the infrared palm image to obtain spectral information;

[0012] Step S3: Extract the high-frequency part and the low-frequency part from the spectral information;

[0013] Step S4: Perform an inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; perform an inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image;

[0014] Step S5: Use a liveness recognition model to perform liveness recognition on the high-frequency infrared palm image and the high-frequency infrared palm image.

[0015] Optionally, in the infrared palm liveness detection method, it is characterized in that Step S2 includes:

[0016] Step S21: Perform a two-dimensional Fourier transform on the infrared palm image to obtain a complex matrix;

[0017] Step S22: Move the low-frequency part in the spectral information to the center of the spectrum.

[0018] Optionally, in the infrared palm liveness detection method, it is characterized in that Step S3 includes:

[0019] Step S31: Use a high-pass filter to filter out the low-frequency part to obtain the high-frequency part;

[0020] Step S32: Use a low-pass filter to filter out the high-frequency part to obtain the low-frequency part.

[0021] Optionally, in the infrared palm liveness detection method, it is characterized in that Step S4 includes:

[0022] Step S41: Perform an inverse Fourier transform on the high-frequency part to reconstruct the high-frequency infrared palm image;

[0023] Step S42: Perform an inverse Fourier transform on the low-frequency part to reconstruct the low-frequency infrared palm image.

[0024] Optionally, in the infrared palm liveness detection method, it is characterized in that Step S5 includes:

[0025] Step S51: Extract features from the high-frequency infrared palm image to obtain high-frequency palm features; extract features from the low-frequency infrared palm image to obtain low-frequency palm features;

[0026] Step S52: Fuse the high-frequency palm features and the low-frequency palm features to obtain comprehensive features;

[0027] Step S53: Input the comprehensive features into the live body recognition model to obtain the live body recognition result.

[0028] Optionally, in the infrared palm live body detection method, when the live body recognition model is trained, the infrared palm image is used for supervision.

[0029] Optionally, in the infrared palm live body detection method, when the live body recognition model is trained, by setting different parameters, different high-frequency parts and low-frequency parts are obtained to improve the adaptability of the live body recognition model.

[0030] In a second aspect, the present invention provides an infrared palm live body detection system for implementing the infrared palm live body detection method described in any one of the foregoing items, which is characterized by including:

[0031] An acquisition module for acquiring an infrared palm image;

[0032] A transformation module for performing a Fourier transform on the infrared palm image to obtain spectral information;

[0033] An extraction module for extracting the high-frequency part and the low-frequency part from the spectral information;

[0034] An inverse transformation module for performing an inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; and performing an inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image;

[0035] A recognition module for performing live body recognition on the high-frequency infrared palm image and the high-frequency infrared palm image by using a live body recognition model.

[0036] In a third aspect, the present invention provides an infrared palm live body detection device, which is characterized by including:

[0037] A processor;

[0038] A memory storing executable instructions of the processor;

[0039] Wherein, the processor is configured to execute the steps of the infrared palm live body detection method described in any one of the foregoing items by executing the executable instructions.

[0040] In a fourth aspect, the present invention provides a computer-readable storage medium for storing a program, which is characterized in that when the program is executed, the steps of the infrared palm live body detection method described in any one of the foregoing items are implemented.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention not only extracts the high-frequency part (reflecting the texture and detail features of the palm), but also extracts the low-frequency part (reflecting the overall structure and temperature distribution of the palm). The high-frequency features can effectively identify the texture details of the palm, while the low-frequency features can capture the physiological characteristics of the living palm, such as temperature distribution and blood flow. The fusion of such multi-band features makes the living body detection more comprehensive and accurate. The combination of high-frequency and low-frequency features can better cope with the changes of palm images in different environments. For example, in scenes with uneven illumination or complex backgrounds, the low-frequency features can provide stable supplementary information, thereby improving the robustness of the detection.

[0043] The present invention uses Fourier transform to convert the image from the spatial domain to the frequency domain, which can effectively separate the high-frequency and low-frequency information. This frequency-domain analysis method can avoid the problems of noise and detail interference in traditional spatial-domain analysis, making the feature extraction clearer and more stable. Fourier transform and inverse transform are mature mathematical tools, and the calculation process is efficient and easy to implement. Through the Fourier transform algorithm, the extraction and reconstruction of spectral information can be quickly completed, which is suitable for the application scenarios of real-time living body detection.

[0044] The present invention adopts infrared imaging technology to capture the thermal radiation and blood flow conditions of the palm, and these physiological features are the key to living body detection. Compared with visible light imaging, infrared imaging is less dependent on environmental illumination and is difficult to be forged. The present invention can effectively prevent spoofing attacks through means such as photos, videos or molds. The forged palm cannot reproduce the temperature distribution and blood flow characteristics of the living palm, thereby greatly improving the security of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts. By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, objectives and advantages of the present invention will become more obvious:

[0046] Figure 1 It is a step flowchart of an infrared palm living body detection method in an embodiment of the present invention;

[0047] Figure 2 It is a schematic diagram of a high-frequency image in an embodiment of the present invention;

[0048] Figure 3 It is a schematic diagram of a spectrum distribution in an embodiment of the present invention;

[0049] Figure 4 This is a flowchart of steps for obtaining spectral information in an embodiment of the present invention;

[0050] Figure 5 This is a flowchart of steps for extracting high-frequency and low-frequency parts in an embodiment of the present invention;

[0051] Figure 6 This is a flowchart of steps for obtaining high-frequency infrared palm images and low-frequency infrared palm images in an embodiment of the present invention;

[0052] Figure 7 This is a flowchart of steps for performing liveness detection on high-frequency infrared palm images and low-frequency infrared palm images in an embodiment of the present invention;

[0053] Figure 8 This is a schematic structural diagram of an infrared palm liveness detection system in an embodiment of the present invention;

[0054] Figure 9 This is a schematic structural diagram of an infrared palm liveness detection device in an embodiment of the present invention; and

[0055] Figure 10 This is a schematic structural diagram of a computer-readable storage medium in an embodiment of the present invention. Detailed Embodiments

[0056] The present invention will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those of ordinary skill in the art can make several modifications and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0057] The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here, for example, can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0058] An infrared palm liveness detection method provided by an embodiment of the present invention aims to solve the problems existing in the prior art.

[0059] The technical solutions of the present invention and how the technical solutions of the present application solve the above technical problems will be described in detail below with specific embodiments. These several specific embodiments below can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present invention will be described below with reference to the accompanying drawings.

[0060] The present invention obtains an infrared palm image, performs Fourier transform on it to obtain spectral information, extracts the high-frequency part and the low-frequency part respectively, then obtains a high-frequency infrared palm image and a low-frequency infrared palm image through inverse Fourier transform, and finally uses a living body recognition model to complete the living body recognition of the palm based on the high-frequency infrared palm image and the low-frequency infrared palm image, having the advantages of device generality, anti-interference, effective feature extraction, strong learning ability, scalability, high security and good reliability.

[0061] Figure 1 It is a flowchart of the steps of an infrared palm living body detection method in an embodiment of the present invention. As Figure 1 shown, the steps of an infrared palm living body detection method in an embodiment of the present invention include:

[0062] Step S1: Obtain an infrared palm image.

[0063] In this step, a professional infrared imaging device (such as an infrared thermal imager or an infrared camera) is used to ensure that the device can capture the thermal radiation information of the palm. The device should have high resolution, high sensitivity and stable temperature measurement ability. The collected infrared image may contain noise or uneven background temperature. Preprocess the image, including:

[0064] Denosing: Use median filtering or Gaussian filtering to remove the noise in the image.

[0065] Normalization: Normalize the image pixel values to the range of [0, 1] for subsequent processing.

[0066] Cropping and alignment: Crop the image to retain the palm area and align the direction of the palm to ensure the consistency of the image.

[0067] Step S2: Perform Fourier transform on the infrared palm image to obtain spectral information.

[0068] In this step, perform two-dimensional Fourier transform on the grayscale infrared image to convert it from the spatial domain to the frequency domain. The formula for Fourier transform is:

[0069]

[0070] where f(x, y) is the pixel value of the image, F(u, v) is the spectral information, and M and N are the width and height of the image.

[0071] Figure 2 Images of different spectra of the palm are shown. In Figure 2 , "Orig" represents the original image, "lowfrerquency" represents the low-spectrum image, and "high frequency" represents the high-frequency image. It can be seen from the figure that the high-frequency part obviously has characteristics different from those of the low-spectrum part and the original image.

[0072] Figure 3 is the spectral distribution. Figure 3 In, "Low Frequency" represents the spectral distribution of the low-frequency part, and "HighFrequency" represents the spectral distribution of the high-frequency part. From Figure 3 it can be seen that the distributions of the two are significantly different and can be used for live body recognition.

[0073] Step S3: Extract the high-frequency part and the low-frequency part from the spectral information.

[0074] In this step, a high-pass filter is used to extract the high-frequency part. An ideal high-pass filter, a Butterworth high-pass filter, or a Gaussian high-pass filter can be selected. The cut-off frequency of the high-pass filter can be set according to actual needs, and usually a relatively high cut-off frequency is selected to retain more high-frequency details. A low-pass filter is used to extract the low-frequency part. An ideal low-pass filter, a Gaussian low-pass filter, or a Butterworth low-pass filter can be selected. The cut-off frequency of the low-pass filter should be relatively low to retain the overall structure and temperature distribution of the image. Apply the high-pass filter to the spectral information to filter out the low-frequency part and only retain the high-frequency information. Apply the low-pass filter to the spectral information to filter out the high-frequency part and only retain the low-frequency information.

[0075] Step S4: Perform an inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; perform an inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image.

[0076] In this step, perform an inverse Fourier transform on the extracted high-frequency part to reconstruct the high-frequency infrared palm image. The inverse transform formula is:

[0077]

[0078] where F high (u, v) is the high-frequency spectral information.

[0079] Low-frequency inverse transform: Perform an inverse Fourier transform on the extracted low-frequency part to reconstruct the low-frequency infrared palm image. The inverse transform formula is:

[0080]

[0081] where Flow (u, v) is the low-frequency spectrum information.

[0082] Perform an absolute value operation (take the modulus) on the inverse-transformed image to eliminate the complex part and obtain the final high-frequency and low-frequency images.

[0083] Step S5: Use the live body recognition model to perform live body recognition on the high-frequency infrared palm image and the high-frequency infrared palm image.

[0084] In this step, extract texture features from the high-frequency infrared palm image, such as using the gray-level co-occurrence matrix (GLCM), local binary pattern (LBP), or edge detection method. Extract temperature distribution features from the low-frequency infrared palm image, such as calculating statistical features such as temperature gradient, mean, and variance. Fuse the features of the high-frequency and low-frequency images to form a comprehensive feature vector. The fusion method can be simple splicing or weighted fusion, depending on the needs of the model.

[0085] The training process of the live body recognition model is as follows:

[0086] Select a model: Select a suitable machine learning or deep learning model for live body recognition, such as a convolutional neural network (CNN), support vector machine (SVM), or random forest.

[0087] Data preparation: Prepare a large number of live and non-live samples for training and validating the model. The samples should cover different palm types (such as different skin colors, ages, and health statuses).

[0088] Training process: Use the labeled data to train the model and optimize the model's parameters to improve accuracy and robustness.

[0089] The live body detection process is as follows:

[0090] Input image: Input the high-frequency and low-frequency infrared palm images to be detected into the trained live body recognition model.

[0091] Output result: The model will output a probability value or classification result to determine whether the palm is a live body. According to the set threshold (such as 0.5), determine the final live body detection result.

[0092] Figure 4 This is the flowchart of the steps for obtaining spectrum information in an embodiment of the present invention. As Figure 4 shown, the steps for obtaining spectrum information in an embodiment of the present invention include:

[0093] Step S21: Perform a two-dimensional Fourier transform on the infrared palm image to obtain a complex matrix.

[0094] In this step, the Fourier transform is a mathematical tool that converts an image from the spatial domain to the frequency domain and can extract the frequency features of the image. The two-dimensional Fourier transform is a commonly used method for processing image data, and its result is a complex matrix that contains the amplitude and phase information of the image.

[0095] Assume that the input infrared palm image is a two-dimensional grayscale image f(x, y), where x and y represent the horizontal and vertical coordinates of the image respectively, and the image size is M×N. The mathematical expression of the two-dimensional Fourier transform is:

[0096]

[0097] Where:

[0098] F(u, v) is the complex matrix in the frequency domain, representing the spectrum information of the image.

[0099] f(x, y) is the image pixel value in the spatial domain.

[0100] M and N are the width and height of the image respectively.

[0101] u and v are the coordinates in the frequency domain.

[0102] The result of the two-dimensional Fourier transform is a complex matrix F(u, v), where each element contains amplitude and phase information. This complex matrix is the basis for subsequent spectrum analysis.

[0103] Step S22: Move the low-frequency part in the spectrum information to the center of the spectrum.

[0104] In this step, since in the frequency domain, the low-frequency part is usually concentrated in the four corners of the spectrum, while the high-frequency part is distributed in the central region of the spectrum. For the convenience of analysis and processing, it is usually necessary to move the low-frequency part to the center of the spectrum, and this operation is called spectrum centering. Spectrum centering facilitates subsequent filtering operations (such as high-pass filtering and low-pass filtering), makes the distribution of the spectrum more intuitive, and is convenient for visualization and analysis.

[0105] Spectrum centering can be achieved by performing a shifting operation on the complex matrix F(u, v). The shifting formula is:

[0106] F′(u, v) = F(u, v)·(-1) u+v

[0107] Or use a more intuitive shifting method:

[0108] F′(u, v) = F(u - M / 2, v - N / 2)

[0109] Where F′(u, v) is the centered spectrum.

[0110] After the spectrum is centered, the amplitude of the spectrum can be visualized to better understand the frequency distribution of the image. Usually, the amplitude of the spectrum is calculated and the logarithm is taken to enhance the contrast.

[0111] Figure 5 This is a flowchart of the steps for extracting the high-frequency part and the low-frequency part in an embodiment of the present invention. As Figure 5 shown, the steps for extracting the high-frequency part and the low-frequency part in an embodiment of the present invention include:

[0112] Step S31: Use a high-pass filter to filter out the low-frequency part and obtain the high-frequency part.

[0113] In this step, the role of the high-pass filter (High-Pass Filter, HPF) is to retain the high-frequency information of the image while suppressing the low-frequency information. The high-frequency part usually contains the details, textures, and edge information of the image, and these features are very important for texture analysis in live detection.

[0114] Commonly used high-pass filters include the ideal high-pass filter, Butterworth high-pass filter, and Gaussian high-pass filter.

[0115] Ideal high-pass filter: In the frequency domain, the ideal high-pass filter is a circular mask with zero in the center and one at the edge. Its cut-off frequency D 0 determines the degree to which the low-frequency part is suppressed.

[0116] Butterworth high-pass filter: The frequency response of this filter is smoother, avoiding the ringing effect that may be introduced by the ideal filter. Its formula is:

[0117]

[0118] where D(u, v) is the distance from the point in the frequency domain to the center, and n is the order of the filter.

[0119] Gaussian high-pass filter: Its frequency response is similar to the Gaussian distribution, and the formula is:

[0120]

[0121] where σ is the standard deviation, which determines the cut-off frequency of the filter.

[0122] When applying the high-pass filter, the designed high-pass filter H(u, v) is multiplied by the spectrum information F(u, v) to obtain the spectrum information F high (u, v) of the high-frequency part:

[0123] F high (u, v) = H(u, v) × F(u, v)

[0124] The spectrum F after high-pass filteringhigh (u, v) mainly contains high-frequency information and can be used for subsequent high-frequency feature extraction.

[0125] Step S32: Use a low-pass filter to filter out the high-frequency part and obtain the low-frequency part.

[0126] In this step, the role of the low-pass filter (Low-Pass Filter, LPF) is to retain the low-frequency information of the image while suppressing the high-frequency information. The low-frequency part usually contains the overall structure and temperature distribution of the image, and these features are very important for physiological feature analysis in live detection.

[0127] The low-pass filter mainly includes: ideal low-pass filter, Butterworth low-pass filter, and Gaussian low-pass filter.

[0128] Ideal low-pass filter: In the frequency domain, the ideal low-pass filter is a circular mask with a center of one and an edge of zero. Its cut-off frequency D 0 determines the degree to which the high-frequency part is suppressed.

[0129] Butterworth low-pass filter: Its frequency response is smoother, avoiding the ringing effect that may be introduced by the ideal filter. Its formula is:

[0130]

[0131] Gaussian low-pass filter: Its frequency response is similar to the Gaussian distribution, and the formula is:

[0132]

[0133] Multiply the designed low-pass filter H(u, v) by the spectral information F(u, v) to obtain the spectral information F low (u, v):

[0134] F low (u, v) = H(u, v) × F(u, v)

[0135] The spectrum F low (u, v) mainly contains low-frequency information and can be used for subsequent low-frequency feature extraction.

[0136] Figure 6 This is a flowchart of the steps for obtaining a high-frequency infrared palm image and a low-frequency infrared palm image in an embodiment of the present invention. As Figure 6 shown, the steps for obtaining a high-frequency infrared palm image and a low-frequency infrared palm image in an embodiment of the present invention include:

[0137] Step S41: Perform an inverse Fourier transform on the high-frequency part to reconstruct the high-frequency infrared palm image.

[0138] In this step, the inverse Fourier transform is the process of converting the data in the frequency domain back to the spatial domain. For the high-frequency part, the image after the inverse transform will highlight the details, textures, and edge information of the palm, which are crucial for texture analysis in liveness detection. The spectral information F high (u, v) of the high-frequency part is the spectrum processed by the high-pass filter.

[0139] The formula for the two-dimensional inverse Fourier transform is:

[0140]

[0141] Where:

[0142] f high (x, y) is the reconstructed high-frequency image.

[0143] F high (u, v) is the spectrum of the high-frequency part.

[0144] M and N are the width and height of the image.

[0145] The img_high after the inverse transform is the reconstructed high-frequency infrared palm image, which mainly contains the texture and detail information of the palm.

[0146] Step S42: Perform an inverse Fourier transform on the low-frequency part to reconstruct the low-frequency infrared palm image.

[0147] In this step, for the low-frequency part, the image after the inverse Fourier transform will highlight the overall structure and temperature distribution of the palm. These features reflect the physiological characteristics of the palm and are very important for physiological feature analysis in liveness detection. The spectral information F low (u, v) of the low-frequency part is the spectrum processed by the low-pass filter.

[0148] Similar to the high-frequency part, the inverse transform formula for the low-frequency part is:

[0149]

[0150] The img_low after the inverse transform is the reconstructed low-frequency infrared palm image, which mainly contains the overall structure and temperature distribution information of the palm.

[0151] Figure 7 This is the flowchart of the steps for liveness recognition of high-frequency infrared palm images and low-frequency infrared palm images in an embodiment of the present invention. As Figure 7 shown, the steps for liveness recognition of high-frequency infrared palm images and low-frequency infrared palm images in an embodiment of the present invention include:

[0152] Step S51: Extract features from the high-frequency infrared palm image to obtain high-frequency palm features; extract features from the low-frequency infrared palm image to obtain low-frequency palm features.

[0153] In this step, features that can characterize the differences between live and non-live are extracted from the high-frequency and low-frequency images respectively.

[0154] High-frequency images mainly contain the texture and detail information of the palm and are suitable for extracting texture features. Calculate the texture features of the image through the Gray-Level Co-Occurrence Matrix (GLCM), such as contrast, correlation, energy, and homogeneity. Extract the local texture features of the image through Local Binary Pattern (LBP), which has a certain robustness to illumination changes. Detect edges through edge detection, such as using the Sobel or Canny algorithm to detect edges, and extract edge intensity and direction information.

[0155] Low-frequency images mainly contain the overall structure and temperature distribution information of the palm and are suitable for extracting statistical features. Calculate the global temperature mean and variance of the image through the temperature mean and variance. Calculate the temperature change rate through the temperature gradient to reflect the blood flow situation. Calculate the statistical features for different regions of the image (such as the palm center, finger roots) respectively.

[0156] The high-frequency feature vector high_features and the low-frequency feature vector low_features represent the features of the high-frequency and low-frequency images respectively.

[0157] Step S52: Fuse the high-frequency palm features and the low-frequency palm features to obtain comprehensive features.

[0158] In this step, the high-frequency and low-frequency features are fused to form a comprehensive feature vector for subsequent live body recognition. When performing feature fusion, simple concatenation, weighted fusion, or a feature selection algorithm can be used. Simple concatenation is to directly concatenate the high-frequency feature vector and the low-frequency feature vector into a long vector. Weighted fusion is to assign different weights to the high-frequency and low-frequency features according to the importance of the features, and then perform weighted summation. Feature selection is to use a feature selection algorithm (such as Principal Component Analysis PCA) to reduce the dimension and optimize the fused features. The comprehensive feature vector combined_features contains the feature information of the high-frequency and low-frequency images and can more comprehensively characterize the live body features of the palm.

[0159] Step S53: Input the comprehensive features into the live body recognition model to obtain the live body recognition result.

[0160] In this step, the fused features are input into a pre-trained liveness recognition model, which will output the recognition result of liveness or non-liveness. The liveness recognition model can be a deep learning model, a machine learning model, or other models. For a deep learning model, it can be a Convolutional Neural Network (CNN), which can automatically learn the combination relationship of features and is suitable for processing complex image data. For traditional machine learning models, it can be a Support Vector Machine (SVM), Random Forest, or k-Nearest Neighbor (k-NN), which are suitable for situations with lower feature dimensions.

[0161] When training the model, a large number of liveness and non-liveness samples are used to train the model. The samples should cover different palm types (such as different skin colors, ages, and health states). During the training process, the parameters of the model are optimized to improve accuracy and robustness.

[0162] In the model inference stage, the fused feature vector is input into the model, and the model will output a probability value or a classification result. According to the set threshold (such as 0.5), it is determined whether the palm is a live body. The classification result and confidence level output by the model are used to determine whether the palm is a live body.

[0163] This embodiment can make full use of high-frequency and low-frequency information to achieve efficient and accurate infrared palm liveness detection.

[0164] In some embodiments, when training the liveness recognition model, the infrared palm image is used for supervision. In this embodiment, another HEAD is used for liveness judgment. Using the original infrared palm image for supervision can ensure that the deviation during the model training process will not be too large, thus ensuring the training effect.

[0165] In some embodiments, when training the liveness recognition model, by setting different parameters, different high-frequency parts and low-frequency parts are obtained to improve the adaptability of the liveness recognition model. During the training process of the liveness recognition model, by adjusting the parameters of the high-frequency and low-frequency parts extracted by Fourier transform, the adaptability of the model can be significantly improved. In Fourier transform, the high-frequency part mainly reflects the details and textures of the image, while the low-frequency part reflects the overall structure and temperature distribution. By adjusting the parameters of the filter (such as the cut-off frequency, filter type, etc.), the extraction effect of the high-frequency and low-frequency parts can be changed to adapt to different liveness detection requirements.

[0166] When adjusting the parameters of the high-pass filter, by changing the cut-off frequency D 0 , the amount of high-frequency information retained can be controlled. A higher cut-off frequency will retain more details but may introduce noise; a lower cut-off frequency will filter out more details.

[0167] When adjusting the parameters of the low-pass filter, similarly by changing the cut-off frequency D0 , the amount of low-frequency information retained can be controlled. A lower cut-off frequency will retain more overall structural information but may lose some details.

[0168] When selecting the filter type, different types of filters (such as ideal filters, Gaussian filters, Butterworth filters) can be selected, and each filter has a different impact on image features.

[0169] During the training process, multiple data sources and feature types can be combined to enhance the model's representation ability. For example, while combining high-frequency and low-frequency features, introducing data from other modalities (such as visible light images, depth information, etc.) can further improve the model's robustness. For instance, combining high-frequency and low-frequency features with visible light image features, depth information, etc. to form a multi-modal feature vector. The training strategy of the model can also be dynamically adjusted according to the distribution of the training data, such as optimizing the model's adaptability through adversarial learning or clustering algorithms.

[0170] By adjusting the parameters during the training process (such as learning rate, batch size, loss function, etc.), the performance of the model can be optimized to better adapt to different scenarios. For example, multiple loss functions (such as cosine similarity loss, classification loss, etc.) can be used to optimize the training of the model. The generalization ability of the model can also be enhanced through adversarial learning by iteratively training with different data subsets. According to the hardware computing power of the deployment device, the structure and parameters of the model are optimized to ensure the efficiency and accuracy of the model in practical applications.

[0171] In this embodiment, by adjusting the parameters of the Fourier transform, combining multi-modal data, and optimizing the training strategy, the adaptability of the live body recognition model can be significantly improved. It can not only make full use of the advantages of high-frequency and low-frequency features but also adapt to different application scenarios through flexible training strategies.

[0172] Figure 8 It is a schematic structural diagram of an infrared palm live body detection system in an embodiment of the present invention. As Figure 8 shown, an infrared palm live body detection system in an embodiment of the present invention includes:

[0173] An acquisition module, configured to acquire an infrared palm image;

[0174] A transformation module, configured to perform a Fourier transform on the infrared palm image to obtain spectral information;

[0175] An extraction module, configured to extract the high-frequency part and the low-frequency part from the spectral information;

[0176] An inverse transformation module, configured to perform an inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; and perform an inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image;

[0177] An identification module for performing liveness identification on the high-frequency infrared palm image and the high-frequency infrared palm image by using a liveness identification model.

[0178] Specifically, the acquisition module is responsible for collecting infrared palm images, using an infrared imaging device (such as an infrared thermal imager or an infrared camera) to obtain the thermal radiation image of the palm. This module needs to ensure that the image is clear and complete, and perform preliminary preprocessing on the image (such as denoising, normalization, etc.).

[0179] The transformation module performs Fourier transform on the acquired infrared palm image, converts it from the spatial domain to the frequency domain, and obtains spectral information. The transformation module uses a two-dimensional Fourier transform algorithm (such as the fast Fourier transform, FFT) to process the image data to obtain spectral information. This module also performs centering processing on the spectrum for subsequent analysis.

[0180] The extraction module extracts the high-frequency part and the low-frequency part from the spectral information. The extraction module uses a high-pass filter and a low-pass filter to filter the spectral information respectively. The high-pass filter extracts the high-frequency part, and the low-pass filter extracts the low-frequency part. The parameters of these filters (such as the cut-off frequency) can be adjusted according to actual needs.

[0181] The inverse transformation module performs inverse Fourier transform on the extracted high-frequency and low-frequency parts to reconstruct the high-frequency and low-frequency infrared palm images. The inverse transformation module performs two-dimensional inverse Fourier transform on the high-frequency and low-frequency spectra respectively to obtain the reconstructed high-frequency and low-frequency images. These images respectively reflect the details and the overall structure of the palm.

[0182] The identification module uses a liveness identification model to perform liveness identification on the high-frequency and low-frequency infrared palm images. The identification module extracts features from the high-frequency and low-frequency images, and after fusing these features, inputs them into a pre-trained liveness identification model. The model outputs the identification result of liveness or non-liveness.

[0183] In this embodiment, by acquiring an infrared palm image, performing Fourier transform on it to obtain spectral information and respectively extracting the high-frequency part and the low-frequency part, then performing inverse Fourier transform to obtain the high-frequency infrared palm image and the low-frequency infrared palm image, and finally using a liveness identification model to complete the liveness identification of the palm based on the high-frequency infrared palm image and the low-frequency infrared palm image, it has the advantages of general equipment, anti-interference, effective feature extraction, strong learning ability, scalability, high security and good reliability.

[0184] In an embodiment of the present invention, there is also provided an infrared palm liveness detection device, including a processor. A memory storing executable instructions of the processor. Wherein, the processor is configured to execute the steps of an infrared palm liveness detection method by executing the executable instructions.

[0185] As described above, in this embodiment, by obtaining an infrared palm image, performing Fourier transform on it to obtain spectral information, and separately extracting the high-frequency part and the low-frequency part, and then performing inverse Fourier transform to obtain a high-frequency infrared palm image and a low-frequency infrared palm image, and finally using a live body recognition model to complete the live body recognition of the palm based on the high-frequency infrared palm image and the low-frequency infrared palm image, it has the advantages of device generality, anti-interference, effective feature extraction, strong learning ability, scalability, high security, and good reliability.

[0186] Those skilled in the art of the present technology can understand that various aspects of the present invention can be implemented as a system, a method, or a program product. Therefore, various aspects of the present invention can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "platform" here.

[0187] Figure 9 It is a schematic structural diagram of an infrared palm live body detection device in an embodiment of the present invention. The following refers to Figure 9 to describe the electronic device 600 according to this embodiment of the present invention. Figure 9 The electronic device 600 shown is only an example and should not impose any restrictions on the functions and usage scope of the embodiments of the present invention.

[0188] As Figure 9 shown, the electronic device 600 is presented in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0189] Among them, the storage unit stores program code, and the program code can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the above part of an infrared palm live body detection method in this specification. For example, the processing unit 610 can execute the steps as Figure 1 shown in.

[0190] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only storage unit (ROM) 6203.

[0191] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205. Such program modules 6205 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a grid environment.

[0192] The bus 630 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus structures.

[0193] The electronic device 600 may also communicate with one or more external devices 700 (such as a keyboard, a pointing device, a Bluetooth device, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or may communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication may be carried out through the input / output (I / O) interface 650. Also, the electronic device 600 may communicate with one or more grids (such as a local area network (LAN), a wide area network (WAN), and / or a public grid, such as the Internet) through the grid adapter 660. The grid adapter 660 may communicate with other modules of the electronic device 600 through the bus 630. It should be understood that although Figure 9 not shown, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0194] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, and the steps of an infrared palm liveness detection method are implemented when the program is executed. In some possible implementation manners, various aspects of the present invention may also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above-mentioned part of the infrared palm liveness detection method of this specification.

[0195] As shown above, in this embodiment, by acquiring an infrared palm image, performing a Fourier transform on it to obtain spectral information and separately extracting the high-frequency part and the low-frequency part, and then performing an inverse Fourier transform to obtain a high-frequency infrared palm image and a low-frequency infrared palm image, and finally using a liveness recognition model to complete the liveness recognition of the palm based on the high-frequency infrared palm image and the low-frequency infrared palm image, it has the advantages of device generality, anti-interference, effective feature extraction, strong learning ability, scalability, high security, and good reliability.

[0196] Figure 10 is a schematic structural diagram of a computer-readable storage medium in an embodiment of the present invention. Refer to Figure 10 As shown, a program product 800 for implementing the above method according to an embodiment of the present invention is described. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device.

[0197] The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0198] The computer-readable storage medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, and this readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted by any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0199] The program code for performing the operations of the present invention can be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user's computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., by using an Internet service provider to connect through the Internet).

[0200] In this embodiment, by obtaining an infrared palm image, performing Fourier transform on it to obtain spectral information and separately extracting the high-frequency part and the low-frequency part, then performing inverse Fourier transform to obtain a high-frequency infrared palm image and a low-frequency infrared palm image, and finally using a live body recognition model to complete the live body recognition of the palm based on the high-frequency infrared palm image and the low-frequency infrared palm image, it has the advantages of device generality, anti-interference, effective feature extraction, strong learning ability, scalability, high security and good reliability.

[0201] The various embodiments in this specification are described in a progressive manner, and the key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0202] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various deformations or modifications within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. An infrared palm liveness detection method, characterized in that: include: Step S1: Acquire infrared palm image; Step S2: Perform Fourier transform on the infrared palm image to obtain spectrum information; Step S3: extracting the high frequency part and the low frequency part in the spectrum information; Step S4: performing inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; performing inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image; Step S5: Performing liveness recognition on the high-frequency infrared palm image and the high-frequency infrared palm image using a liveness recognition model.

2. The infrared palm liveness detection method according to claim 1, characterized in that: Step S2 includes: Step S21: performing a two-dimensional Fourier transform on the infrared palm image to obtain a complex matrix; Step S22: Move the low-frequency part in the spectrum information to the center of the spectrum.

3. The infrared palm liveness detection method according to claim 1, characterized in that: Step S3 includes: Step S31: Use a high-pass filter to filter out the low-frequency part to obtain the high-frequency part; Step S32: Use a low-pass filter to filter out the high-frequency part to obtain the low-frequency part.

4. The infrared palm liveness detection method according to claim 1, characterized in that: Step S4 includes: Step S41: performing inverse Fourier transform on the high-frequency part to reconstruct a high-frequency infrared palm image; Step S42: performing inverse Fourier transform on the low-frequency part to reconstruct a low-frequency infrared palm image.

5. The infrared palm liveness detection method according to claim 1, characterized in that: Step S5 includes: Step S51: extracting features from the high-frequency infrared palm image to obtain high-frequency palm features; extracting features from the low-frequency infrared palm image to obtain low-frequency palm features; Step S52: fusing the high-frequency palm feature and the low-frequency palm feature to obtain a comprehensive feature; Step S53: input the comprehensive features into a liveness recognition model to obtain a liveness recognition result.

6. The infrared palm liveness detection method according to claim 1, characterized in that: The living body recognition model uses the infrared palm image as supervision during training.

7. The infrared palm liveness detection method according to claim 1, characterized in that: When the liveness recognition model is trained, different high-frequency parts and low-frequency parts are obtained by setting different parameters, thereby improving the adaptability of the liveness recognition model.

8. An infrared palm liveness detection system, used to implement the infrared palm liveness detection method according to any one of claims 1 to 7, characterized in that: include: An acquisition module, used for acquiring infrared palm images; A transformation module, used for performing Fourier transformation on the infrared palm image to obtain spectrum information; An extraction module, used to extract the high frequency part and the low frequency part in the spectrum information; An inverse transform module, used for performing inverse Fourier transform on the high-frequency part to obtain a high-frequency infrared palm image; Performing inverse Fourier transform on the low-frequency part to obtain a low-frequency infrared palm image; The recognition module is used to perform liveness recognition on the high-frequency infrared palm image and the high-frequency infrared palm image using a liveness recognition model.

9. An infrared palm liveness detection device, characterized in that: include: processor; a memory storing executable instructions of the processor; The processor is configured to execute the steps of the infrared palm liveness detection method according to any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium for storing a program, characterized in that: When the program is executed, the steps of the infrared palm liveness detection method according to any one of claims 1 to 7 are implemented.

Images