Electronic signature method and device, electronic equipment and computer program product

Abstract

The invention discloses an electronic signature method and device, electronic equipment and a computer program product. The method comprises the steps that multi-modal biological characteristics of a target object are collected, and the multi-modal biological characteristics at least comprise biological static characteristics of multiple modalities and biological dynamic characteristics of multiple modalities; fusing the biological static features of the multiple modalities into biological fusion features; distributing a private key to the target object according to the biological fusion feature under the condition that the biological dynamic feature passes the living body detection; the signature information is encrypted by using the private key to obtain an electronic signature, and the signature information is the original text to be signed. According to the invention, the technical problem of poor security of the electronic signature in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of computers, and more specifically, to an electronic signature method, apparatus, electronic device, and computer program product. Background Technology

[0002] With the rapid development of digital office and e-commerce, electronic signature technology, as a key technology to ensure the legal validity and security of electronic documents, has been widely used. Currently, mainstream electronic signature technologies are mainly based on Public Key Infrastructure (PKI) systems, using a combination of digital certificates and hash algorithms to achieve document signing and verification. While this type of technology meets the requirements for the integrity and non-repudiation of electronic documents to a certain extent, it still faces many challenges in practical applications.

[0003] For example, traditional PKI systems rely on Central Certification Authorities (CAs). If a CA server is attacked or has internal management vulnerabilities, digital certificates may be forged or tampered with, which could lead to the risk of signature forgery.

[0004] Furthermore, static digital certificates cannot effectively handle situations where an object's identity is misused. Existing electronic signature processes typically require the object to manually enter a password and insert a USB key, which is cumbersome in mobile work scenarios and fails to meet the need for convenience. Additionally, batch processing efficiency is low for large-scale document signing scenarios.

[0005] There is currently no effective solution to the problem of poor security of electronic signatures in the existing technologies mentioned above. Summary of the Invention

[0006] This invention provides an electronic signature method, apparatus, electronic device, and computer program product to at least solve the technical problem of poor security of electronic signatures in the prior art.

[0007] According to one aspect of the present invention, an electronic signature method is provided, comprising: collecting multimodal biometrics of a target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; fusing the multiple modalities of static biometrics into a biometric fusion feature; when the dynamic biometrics pass liveness detection, allocating a private key to the target object based on the biometric fusion feature; and encrypting signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0008] Optionally, fusing the biostatic features of multiple modalities into a biofusion feature includes: identifying the business scenario in which the multimodal biostatistics are collected, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to the biostatic features of multiple modalities respectively; and performing weighted fusion of the biostatic features of multiple modalities according to the feature weight combination corresponding to the business scenario to obtain the biofusion feature.

[0009] Optionally, before weighted fusing the biostatic features of multiple modalities according to the feature weight combination corresponding to the business scenario to obtain the biofusion feature, the method further includes: determining the recognition confidence of the biostatic feature of each modality, wherein the recognition confidence is used to represent the degree of matching between the biostatic feature and the template features in the preset feature library; determining the biostatic features with recognition confidence lower than a first recognition threshold as features to be adjusted; reducing the corresponding feature weight for the features to be adjusted, and increasing the corresponding feature weight for the biostatic features that do not belong to the features to be adjusted, to obtain the feature weight combination.

[0010] Optionally, reducing the corresponding feature weight for the feature to be adjusted and increasing the corresponding feature weight for the biostatic features that do not belong to the feature to be adjusted, to obtain the feature weight combination includes: counting the number of consecutive historical occurrences of the recognition confidence of the feature to be adjusted being lower than the first recognition confidence threshold; if the number of consecutive historical occurrences is greater than a preset number threshold, reducing the feature weight corresponding to the feature to be adjusted to a preset minimum weight; increasing the corresponding feature weight for the biostatic features whose recognition confidence is greater than a second confidence threshold; and generating the feature weight combination based at least on the preset minimum weight corresponding to the feature to be adjusted and the increased feature weight.

[0011] Optionally, before reducing the corresponding feature weight for the feature to be adjusted and increasing the corresponding feature weight for the biostatic feature that does not belong to the feature to be adjusted to obtain the feature weight combination, the method further includes: detecting and collecting various environmental information of the multimodal biofeature; identifying the environmental information that does not meet the preset environmental conditions as abnormal environmental information; and identifying the biostatic feature whose recognition accuracy is negatively affected by the abnormal environmental information as the feature to be adjusted.

[0012] Optionally, if the biodynamic features pass the liveness detection, before querying the private key pre-allocated to the target object based on the biofusion features, the method further includes: detecting whether the biodynamic features of each modality meet the corresponding dynamic feature verification criteria; and determining that the biodynamic features pass the liveness detection if the biodynamic features of each modality meet the corresponding dynamic feature verification criteria.

[0013] Optionally, if the biodynamic features of each modality meet the corresponding dynamic feature verification criteria, determining that the biodynamic features pass the liveness detection includes: detecting whether the time difference between the acquisition timestamp of the multimodal biofeature and the blockchain timestamp does not exceed a preset time threshold, wherein the blockchain timestamp is obtained by initiating a timestamp query request to the blockchain network, and the timestamp query request is initiated immediately after the biostatic feature acquisition is completed; if the time difference does not exceed the preset time threshold, determining that the biodynamic features pass the liveness detection.

[0014] Optionally, when the biodynamic feature passes the liveness detection, allocating a private key to the target object based on the biofusion feature includes: hashing the biofusion feature to obtain a biometric hash value; verifying the biometric hash value against pre-set biometric constraints to obtain a feature matching value; determining a corresponding feature matching threshold based on a pre-set risk level for the signing information, wherein each risk level has a pre-set feature matching threshold; and allocating a private key to the target object when the feature matching value is greater than the feature matching threshold.

[0015] Optionally, when the feature matching value is greater than the feature matching threshold, the private key allocated to the target object includes: determining identity information matching the biometric fusion feature when the feature matching value is greater than the feature matching threshold, wherein the identity information is pre-configured with identity matching conditions matching the biometric fusion feature; determining the signing permission of the target object based on the identity information, wherein the identity information is also pre-configured with corresponding signing permissions; determining the permission verification conditions for the signing information in a pre-set permission rule, wherein the permission rule pre-sets multiple permission verification conditions and a level matching relationship between each permission verification condition and the signing information; and allocating the private key to the target object when the signing permission meets the permission verification conditions corresponding to the signing information.

[0016] Optionally, encrypting the signing information using the private key to obtain an electronic signature includes: hashing the signing information to obtain a hash value of the signing information; and encrypting the hash value of the signing information using the private key to obtain the electronic signature.

[0017] Optionally, after encrypting the signing information using the private key to obtain an electronic signature, the method further includes: generating signature information based on the electronic signature, wherein the signature information includes at least the electronic signature; and uploading the signature information to a blockchain network.

[0018] Optionally, after uploading the signature information to the blockchain network, the method further includes: hashing the signature information to be verified to obtain a verification hash value; extracting the electronic signature from the signature information on the blockchain network as a signature to be verified; decrypting the signature to be verified using a public key pre-configured for the target object to obtain a decryption hash value; and determining that the signature information to be verified has not been tampered with if the verification hash value matches the decryption hash value.

[0019] According to another aspect of the present invention, an electronic signature device is also provided, comprising: a collection module for collecting multimodal biometrics of a target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; a fusion module for fusing the multiple modalities of static biometrics into a biometric fusion feature; an allocation module for allocating a private key to the target object based on the biometric fusion feature when the dynamic biometrics pass liveness detection; and an encryption module for encrypting signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0020] According to another aspect of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to execute the above-described electronic signature method through the computer program.

[0021] According to another aspect of the present invention, a computer program product is also provided, including computer instructions that, when executed by a processor, implement the steps of the above-described electronic signature method.

[0022] The embodiments described above in this application collect multimodal biometric features of the target object and use a dynamic weight adjustment algorithm to fuse multimodal static features into biometric fusion features. This ensures high-precision biometric recognition even in complex environments. Furthermore, when the biometric dynamic features pass liveness detection, a private key is allocated to the target object based on the biometric fusion features. This ensures that the allocation of the private key is strictly based on real-time and authentic biometric features, greatly improving the security of the private key. It also ensures that the electronic signature encrypted with this private key has high personal attributes and is not easily copied, effectively preventing the forgery of electronic signatures and the impersonation of the target object's identity. This ensures the authenticity and uniqueness of the signing behavior, thereby achieving the technical effect of improving the security of electronic signatures and solving the technical problem of poor security of electronic signatures in the prior art. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0024] Figure 1 This is a flowchart of an electronic signature method according to an embodiment of the present invention;

[0025] Figure 2 This is an illustration of an electronic signature according to an embodiment of the present invention. Figure 1 ;

[0026] Figure 3 This is an illustration of an electronic signature according to an embodiment of the present invention. Figure 2 ;

[0027] Figure 4 This is a schematic diagram of an electronic signature device according to an embodiment of the present invention;

[0028] Figure 5 This is a structural block diagram of a computer terminal according to an embodiment of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] First, some nouns or terms that appear in the description of the embodiments of this application shall be interpreted as follows:

[0032] Blockchain model: refers to a blockchain system used to process data related to electronic signatures.

[0033] Smart contracts are automatically executed contract terms deployed on a blockchain in the form of computer programs.

[0034] Transaction module: In blockchain, a transaction module refers to a data structure that contains specific information (such as the hash value of the signature information, the signature information, the biometric summary of the object, the signature timestamp, etc.).

[0035] Consensus verification: refers to the process by which multiple nodes in a blockchain network verify the validity of a transaction module.

[0036] Blockchain timestamp: refers to the time stamp of an event recorded in a blockchain system.

[0037] Hash value: A fixed-length digital digest obtained by processing data using a hash algorithm.

[0038] Private key and public key: These are a pair of keys used in encryption algorithms. The private key is used to sign data, and the public key is used to verify the signature.

[0039] Dynamic threshold mechanism: This is a mechanism that sets different matching thresholds based on different business risk levels.

[0040] Anomaly Handling Process: This refers to a series of handling measures automatically triggered by the system when anomalies are detected during the signing process (such as too many failed biometric verification attempts, blockchain network anomalies, etc.).

[0041] According to an embodiment of the present invention, an embodiment of an electronic signature method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0042] Figure 1 This is a flowchart of an electronic signature method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0043] Step S102: Collect multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics;

[0044] Step S104: Fuse the biostatic features of multiple modalities into a biofusion feature;

[0045] Step S106: If the biodynamic features pass the liveness detection, allocate a private key to the target object based on the biofusion features.

[0046] Step S108: Encrypt the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0047] The embodiments described above in this application collect multimodal biometric features of the target object and use a dynamic weight adjustment algorithm to fuse multimodal static features into biometric fusion features. This ensures high-precision biometric recognition even in complex environments. Furthermore, when the biometric dynamic features pass liveness detection, a private key is allocated to the target object based on the biometric fusion features. This ensures that the allocation of the private key is strictly based on real-time and authentic biometric features, greatly improving the security of the private key. It also ensures that the electronic signature encrypted with this private key has high personal attributes and is not easily copied, effectively preventing the forgery of electronic signatures and the impersonation of the target object's identity. This ensures the authenticity and uniqueness of the signing behavior, thereby achieving the technical effect of improving the security of electronic signatures and solving the technical problem of poor security of electronic signatures in the prior art.

[0048] In step S102 above, the multimodal biometrics are collected after authorization from the target object.

[0049] Optionally, collecting the multimodal biometrics of the target object includes: detecting the authorization confirmation instruction of the target object; and, in response to the authorization confirmation instruction, collecting the multimodal biometrics of the target object.

[0050] In step S102 above, the multimodal biometrics include: static biometrics for determining identity information and dynamic biometrics for performing liveness detection, wherein both the static and dynamic biometrics have multiple modalities.

[0051] In step S102 above, the biostatic features of multiple modalities include: fingerprint features, facial features and iris features.

[0052] As an optional example, fingerprint feature acquisition includes: acquiring fingerprint images of the target object using an optical fingerprint sensor or a capacitive fingerprint sensor, wherein the acquisition resolution is not less than 500 dpi; judging the acquired image based on image analysis algorithms to obtain the acquired fingerprint result (i.e., acquired features); if the acquired fingerprint result has factors that may affect the accuracy of recognition, such as blurriness, distortion, dirt, or incorrect finger placement, a clear prompt is issued to the target object, guiding the target object to re-acquire fingerprints until a clear fingerprint image that meets the requirements is obtained, thereby ensuring the reliability and validity of fingerprint data (i.e., fingerprint features).

[0053] For example, the target uses the optical fingerprint sensor on the phone to capture a fingerprint image; the phone system automatically detects the image quality and prompts the target to adjust the angle of their finger placement, ultimately successfully obtaining a clear fingerprint image.

[0054] As an optional example, facial feature acquisition includes: acquiring facial images of the subject through a camera, wherein the subject is required to be in a well-lit environment with no facial obstruction and a proper posture; using a face detection algorithm to locate key facial points (such as the corners of the eyes, the tip of the nose, the corners of the mouth, etc.), and normalizing the image to eliminate the influence of factors such as lighting and posture, thereby obtaining facial features.

[0055] For example, the front-facing camera of a mobile phone can be used to capture facial images of a target object. The camera automatically adjusts the lighting to ensure that the image quality meets the requirements, and uses a face detection algorithm to locate key facial points and complete image normalization processing.

[0056] As an optional example, the acquisition of iris features includes: acquiring an image of the iris of the target object using a dedicated iris acquisition device, wherein the object is required to look at the lens during acquisition, the device automatically focuses on the iris region, and a high-resolution iris image (resolution not less than 720p) is obtained; the iris image is preprocessed with noise reduction, normalization and other methods, and iris texture features are extracted to obtain iris features.

[0057] For example, the target subject uses the iris capture function on their mobile phone, follows the prompts to look at the lens, and the phone automatically focuses and acquires a high-resolution iris image, which is then preprocessed to extract iris texture features.

[0058] In step S104 above, the biostatic features include: a first static feature, a second static feature, and a third static feature with different modalities. The biostatic features of multiple modalities can be fused into a biofusion feature by weighted summation. Specifically, the first feature weight of the first static feature is determined, the second feature weight of the second static feature is determined, and the third feature weight of the third static feature is determined. Then, the first product of the first static feature and the first feature weight, the second product of the second static feature and the second feature weight, and the third product of the third static feature and the third feature weight are superimposed to obtain the biofusion feature. The first feature weight, the second feature weight, and the third feature weight are the feature weight combination of the biostatic features.

[0059] Optionally, the first static feature can be a fingerprint feature, the second static feature can be a facial feature, and the third static feature can be an iris feature.

[0060] In step S106 above, the biometric fusion feature is unique and can represent the identity information of the target object. Therefore, when allocating a private key to the target object, the allocation can be made based on the biometric fusion feature.

[0061] Optionally, a pre-set server records the correspondence between the target object and the private key. The correspondence between the target object and the private key is established based on the unique identity information of the target object. After the pre-set server identifies the biometric fusion characteristics of the target object, it can determine the identity information of the target object based on the biometric fusion characteristics and generate or query the private key of the target object based on the identity information.

[0062] In the above embodiments of this application, since the private key is directly associated with the biometric fusion feature, the signing information is encrypted using a private key allocated based on the biometric fusion feature of the target object, thus ensuring the integrity of the original document.

[0063] In step S106 above, before allocating a private key to the target object, it is necessary to perform a liveness detection based on the target object's biodynamic characteristics. Only if the liveness detection is passed can the private key be allocated based on the target object's static biodynamic characteristics.

[0064] Optionally, the biodynamic features include at least one of the following: a first dynamic feature, a second dynamic feature, and a third dynamic feature with different modalities; for example, changes in the pressure of fingerprint pressing, micro-movements of facial expressions, and subtle tremors of the iris.

[0065] It should be noted that biodynamic features need to correspond to biostatic features; for example, when the biostatic feature is a fingerprint feature, the biodynamic feature is the fingerprint dynamic feature, such as the change in the pressure of the fingerprint; when the biostatic feature is a facial feature, the biodynamic feature is the facial dynamic feature, such as the micro-movements of facial expressions; when the biostatic feature is a iris feature, the biodynamic feature is the iris dynamic feature, such as the subtle tremors of the iris.

[0066] Optionally, the acquisition of fingerprint dynamic features includes: processing the acquired pressure data using a time series analysis algorithm to extract the time-domain features (pressure peak, valley, and change period) and frequency-domain features (dominant frequency and harmonic components) of the pressure change, and generating a fingerprint dynamic feature vector.

[0067] Optionally, the collection of facial dynamic features includes: analyzing the collected facial images through a facial action coding model to identify action units of micro-movements such as blinking, opening the mouth, and frowning; judging the continuity and authenticity of micro-movements through a temporal modeling algorithm; and excluding forgery methods such as photos and videos.

[0068] For example, when a person is having their face captured, the system will ask the person to complete a series of micro-movements, such as blinking and opening their mouth. The system will use the FACET model to identify the action units of these micro-movements, and then use the Hidden Markov Model to determine the continuity and authenticity of the micro-movements in order to determine whether the face is a live person.

[0069] Optionally, the acquisition of iris dynamic features includes: tracking iris texture feature points in the iris image sequence using a feature point tracking algorithm, calculating the displacement trajectory and velocity of the feature points, and selecting effective dynamic parameters of subtle iris tremors (tremor amplitude 0.05mm-0.2mm, tremor frequency 1Hz-3Hz) to form an iris dynamic feature set.

[0070] Optionally, biometric liveness detection includes at least one of the following: liveness detection of facial dynamic features through actions such as blinking and opening the mouth; and liveness detection of iris dynamic features through the pupil's response to light.

[0071] As an optional example, for liveness detection of fingerprint dynamic features, the verification standard for passing liveness detection is: the change in pressure must meet the requirement that "the fluctuation amplitude of pressure is not less than 0.2N within 3 consecutive sampling periods", and the trend of pressure change must conform to the natural physiological law of human finger pressing (such as the process of pressure from small to large and then to stable). If it does not meet this standard, it is judged as a failure of dynamic verification, that is, it fails the liveness detection.

[0072] As an optional example, for liveness detection of facial dynamic features, the verification criteria for passing liveness detection are: at least two valid micro-movements (such as one complete blink + one slight upturn of the corner of the mouth) must be completed within 10 seconds, and the duration and amplitude of the micro-movements must conform to the characteristics of real human behavior (such as blinking for 0.2s-0.5s). If no valid micro-movements are completed or the micro-movements are mechanically repetitive, the dynamic verification is deemed to have failed, i.e., the liveness detection is not passed.

[0073] As an optional example, for liveness detection of dynamic features of iris patterns, the verification criteria for passing liveness detection are as follows: the amplitude of the subtle iris tremors must be within the range of 0.05mm-0.2mm, and the tremor frequency must be stable between 1Hz and 3Hz. If the detected tremor amplitude continuously exceeds 0.2mm or the frequency is abnormal (such as below 0.5Hz or above 5Hz), it is judged as a failure of dynamic verification, that is, it fails the liveness detection.

[0074] In step S108 above, the signing information refers to the original document content or a certain representation of the document that needs to be signed, such as the text content of the document, image data, or more commonly, the hash value of the document content, that is, the signing information hash value.

[0075] It should be noted that the signature information hash value is calculated from the signature information using a hash algorithm (such as SHA-256). It is a fixed-length digital digest that uniquely represents the content of the original file. Any modification to the original file will cause a change in the hash value; therefore, the hash value is used to verify the integrity of the file.

[0076] In step S108 above, the electronic signature is obtained by encrypting the signing information (such as the hash value of the signing information) using the signer's (i.e., the target object's) private key. The electronic signature can verify that the signing information was indeed signed by the target object himself and has not been tampered with by others.

[0077] The technical solution of this embodiment can significantly improve the security and reliability of electronic signatures, providing a safer and more reliable solution for document signing in digital office and e-commerce scenarios.

[0078] As an optional embodiment, fusing multiple modal biostatic features into a biofusion feature includes: identifying business scenarios for collecting multimodal biometrics, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to multiple modal biostatic features respectively; and performing weighted fusion of multiple modal biostatic features according to the feature weight combination corresponding to the business scenario to obtain the biofusion feature.

[0079] The embodiments described above dynamically adjust the feature weights corresponding to the static biometric features of each modality in the multimodal biometrics for different business scenarios. That is, the feature weight combination is determined based on the differences in business scenarios, so that the feature weights can be automatically adjusted according to changes in business scenarios, thereby achieving accurate identification of target objects, effectively preventing security vulnerabilities in the biometric identification process, improving the overall security and reliability of the electronic signature system, and solving the technical problem of poor security of electronic signatures in the prior art.

[0080] For example, in mobile office scenarios, the weight of fingerprint is 0.4, the weight of face is 0.3, and the weight of iris is 0.3; while in high-security financial scenarios, the weight of iris increases to 0.5, the weight of fingerprint is 0.3, and the weight of face is 0.2. Based on the feature weight combination corresponding to the business scenario, the static biometric features of fingerprint, face, and iris are weighted and fused to obtain biometric fusion features that adapt to the needs of the scenario.

[0081] Optionally, the business scenarios may include at least: mobile office scenarios (such as outdoors or inside a vehicle), indoor office scenarios (such as those with stable lighting and no obvious obstructions), and high-security financial scenarios (signing large contracts).

[0082] Optionally, for mobile office scenarios (outdoors, inside vehicles), experimental data shows that the fingerprint recognition error rate is 3.2% (affected by hand dirt and pressing angle), the face recognition error rate is 8.5% (affected by lighting and occlusion), and the iris recognition error rate is 2.8% (affected by device stability and subject cooperation). Considering both convenience (fastest fingerprint acquisition speed) and security (lowest iris recognition error rate), setting the fingerprint weight to 0.4, the face weight to 0.3, and the iris weight to 0.3 (i.e., the feature weight combination corresponding to the mobile office scenario) reduces the fusion recognition error rate to below 1.5%.

[0083] Optionally, for indoor office scenarios (stable lighting, no significant obstructions): Experimental data shows that the face recognition error rate is reduced to 1.8%, the fingerprint recognition error rate is 2.5%, and the iris recognition error rate is 2.2%. At this point, increasing the face weight to 0.4, adjusting the fingerprint weight to 0.3, and adjusting the iris weight to 0.3 (which is the feature weight combination corresponding to the indoor office scenario) can further reduce the fusion recognition error rate to below 1.0%.

[0084] Optionally, for high-security financial scenarios (signing large contracts), experimental data shows that iris recognition has the lowest error rate (1.5%), fingerprint recognition has an error rate of 2.0%, and face recognition has an error rate of 3.0%. To prioritize security, the weight of iris recognition is set to 0.5, the weight of fingerprint recognition to 0.3, and the weight of face recognition to 0.2 (which is the feature weight combination corresponding to high-security financial scenarios), and the fusion recognition error rate can be controlled within 0.8%.

[0085] The embodiments described above in this application, by weighted fusion of multimodal biometric static features, can obtain more accurate biometric fusion features. Based on these biometric fusion features, not only is the accuracy of biometric identification improved and the risk of misjudgment that may be caused by single biometric identification reduced, but the flexibility and adaptability of the system are also enhanced. The identification strategy can be dynamically adjusted according to the security requirements of different scenarios, thereby ensuring security while taking into account user experience and ease of operation.

[0086] It should be noted that the biostatic features of each modality are identified using their respective feature recognition models. Each feature recognition model uses recognition confidence to evaluate the recognition results of the biostatic features when recognizing their respective biostatic features.

[0087] Optionally, the feature recognition model is used to extract the static biological features of the target object from the collected biological information of the target object.

[0088] Optionally, the confidence score represents the accuracy of the extracted biostatic features of the target object.

[0089] As an optional example, the recognition confidence assessment includes: using a biometric recognition module (such as a feature recognition model for recognizing static biometric features) to recognize each collected fingerprint, face, and iris feature, and then outputting the corresponding recognition confidence (range 0-1, where 1 indicates a perfect match and 0 indicates a complete mismatch).

[0090] For example, each feature recognition model, when recognizing biostatic features, generates multiple recognition results for biostatic features, along with the confidence level of each feature; then, the final biostatic feature is selected based on the confidence level. For instance, the biostatic feature with the highest confidence level is chosen as the output of the feature recognition model.

[0091] As an optional embodiment, before weighted fusion of biostatic features from multiple modalities based on the feature weight combination corresponding to the business scenario to obtain biofusion features, the method further includes: determining the recognition confidence of the biostatic features of each modality, wherein the recognition confidence is used to represent the degree of matching between the biostatic features and template features in a preset feature library; identifying biostatic features with recognition confidence below a first recognition threshold as features to be adjusted; reducing the corresponding feature weights for the features to be adjusted, and increasing the corresponding feature weights for biostatic features that do not belong to the features to be adjusted, to obtain a feature weight combination.

[0092] In the embodiments described above, the accuracy of biometric static features varies across different business environments and biometric collection environments. The accuracy of each modality's biometric static feature can be represented by its recognition confidence level. Therefore, when fusing multimodal biometric static features, the feature weights corresponding to each biometric static feature can be adjusted based on their recognition confidence levels. For example, the feature weights corresponding to modalities with high recognition confidence can be increased, while the feature weights corresponding to modalities with low recognition confidence can be decreased. This dynamically generates feature weight combinations, ensuring that biometric static features with high matching and strong confidence are prioritized during the fusion of multiple biometric information, thus improving the accuracy and reliability of biometric fusion. Furthermore, this dynamic weight adjustment mechanism allows the electronic signature process to be flexibly adjusted according to the data quality of different modalities' biometric static features, prioritizing the dominant role of biometric information with high recognition confidence and stable matching in the multimodal fusion process. This improves the overall security and recognition rate of electronic signatures, solving the technical problem of poor security in existing electronic signature technologies.

[0093] Optionally, the first recognition threshold can be a preset minimum threshold. After determining the recognition confidence of the biostatic features of each modality, if the recognition confidence is lower than the preset first recognition threshold, it indicates that the reliability of these biostatic features is cross-referenced, and the biostatic feature will be marked as a feature to be adjusted. Then, the feature weight corresponding to the biostatic feature is automatically reduced, while the feature weight of biostatic features with high recognition confidence is increased, thereby dynamically generating a feature weight combination.

[0094] For example, in scenarios where the confidence level of iris recognition is low, the feature weight of iris features will be automatically reduced, while the feature weights of fingerprint and facial features will be increased to compensate for the instability of iris recognition and ensure the accuracy and security of identity verification during the electronic signature process.

[0095] As an optional example, the recognition confidence can be determined based on the matching results between the extracted biostatic features and template features in a preset feature library. The preset feature library may pre-record template features of multiple preset objects. After extracting the biostatic features of the target object, the biostatic feature domain template features can be matched for similarity, and the similarity is the recognition confidence.

[0096] It should be noted that the preset feature library can record template features of multiple preset objects, which must also include the template features of the target object. However, the different collection environments of biostatic features will cause slight differences in the biostatic features of the same object collected each time. That is, the biostatic features of the collected target object will not be exactly the same as the template features of the target object recorded in the preset feature library. Therefore, the biostatic features of the target object collected this time and the template features of the target object can be judged by similarity. This similarity is the recognition confidence.

[0097] Furthermore, even if the biostatic features of the target object collected this time differ from the template features of the target object, the difference is much smaller than the difference between the biostatic features of the target object and the template features of other objects. Therefore, based on the recognition confidence, the static biostatic features used to identify the target object can be accurately obtained.

[0098] The embodiments described above in this application automatically optimize weight allocation by monitoring the matching of biometric static features in real time, ensuring that the generation and verification of electronic signatures can achieve the optimal balance in any business scenario, effectively improving the security of the signing process and the user experience.

[0099] As an optional embodiment, the feature weight combination is obtained by reducing the corresponding feature weight of the feature to be adjusted and increasing the corresponding feature weight of the biostatic features that do not belong to the feature to be adjusted. This includes: counting the number of consecutive historical occurrences of the recognition confidence of the feature to be adjusted being lower than a first recognition threshold; reducing the feature weight of the feature to be adjusted to a preset minimum weight when the number of consecutive historical occurrences is greater than a preset number threshold; increasing the corresponding feature weight of the biostatic features that have a recognition confidence greater than a second confidence threshold; and generating a feature weight combination based at least on the preset minimum weight of the feature to be adjusted and the increased feature weight.

[0100] In the embodiments described above, when the recognition confidence of a biostatic feature of a certain modality is continuously lower than a first recognition threshold, the number of such consecutive occurrences is counted. If the number of such consecutive occurrences exceeds a preset threshold, the feature weight corresponding to the biostatic feature of that modality is automatically reduced to a preset minimum weight. At the same time, for biostatic features with a recognition confidence greater than a second confidence threshold, the feature weight corresponding to the biostatic feature of that modality is increased, generating a new feature weight combination to maintain the overall accuracy and stability of biostatic feature recognition, thereby improving the security and reliability of electronic signatures and solving the technical problem of poor security of electronic signatures in the prior art.

[0101] It should be noted that the historical consecutive counts represent the number of times the recognition confidence of the same modality of biostatic features has been continuously lower than the preset recognition threshold.

[0102] As an optional example, if the recognition confidence of a biostatic feature of a certain modality is <0.6 for three consecutive times, the feature weight corresponding to the biostatic feature of that modality is automatically reduced to 0.1, and the reduced feature weight is allocated to other biostatic features with recognition confidence >0.8 to ensure the reliability of fusion recognition.

[0103] For example, in a scenario where the subject's hand is injured (fingerprint recognition confidence level 0.55): if the algorithm detects consecutive failures in fingerprint recognition, it reduces the feature weight of the fingerprint feature to 0.1, increases the feature weight of the face feature to 0.45, and increases the feature weight of the iris feature to 0.45, thereby strengthening face and iris recognition to ensure successful identity authentication.

[0104] It should be noted that the confidence level of identifying biostatic features of different modalities is affected by environmental information (such as light intensity) of the biological data (i.e., biostatic features) collected for each modality.

[0105] For example, when the light intensity is below the ideal threshold, it may affect the accuracy of facial recognition, thereby reducing the feature weight of the modality where the facial features are located, while increasing the feature weight of fingerprint or iris features, to ensure accurate biometric static feature recognition and reliable electronic signature process even in low-light environments.

[0106] As an optional embodiment, before reducing the corresponding feature weights of the features to be adjusted and increasing the corresponding feature weights of the biostatic features that do not belong to the features to be adjusted, and obtaining the feature weight combination, the method further includes: detecting and collecting various environmental information of multimodal biofeatures; identifying environmental information that does not meet the preset environmental conditions as abnormal environmental information; and identifying biostatic features whose recognition accuracy is negatively affected by abnormal environmental information as features to be adjusted.

[0107] In the above embodiments of this application, the factors affecting the confidence level of biometric static feature recognition can be environmental factors that collect the biometric static feature, such as strong light, extreme temperature, or high humidity. These environmental factors may negatively affect the recognition accuracy of specific biometric features. Therefore, by adjusting the corresponding feature weights based on the recognition confidence level of biometric static features of each modality, the influence of abnormal environmental information on biometric static feature recognition can be dynamically analyzed, and biometric features with significantly reduced recognition accuracy can be identified as features to be adjusted. For example, excessively strong or weak light can affect the recognition of facial features, and high humidity may reduce the accuracy of fingerprint recognition. By enhancing the perception and adaptability to environmental factors, the feature weights that are more affected by abnormal environments are automatically reduced, while the feature weights that are not affected by abnormal environments or are less affected are increased. This constructs a more reasonable feature weight combination that is more adapted to the current environment, thereby not only improving the robustness of biometric recognition in complex environments and effectively enhancing the accuracy of identity authentication and the unforgeability of electronic signatures, but also ensuring the security and reliability of the electronic signature system, solving the technical problem of poor security of electronic signatures in the prior art.

[0108] Optionally, the environmental information can be detected in real time by built-in sensors, including various environmental information such as light intensity, ambient temperature and humidity; preset environmental conditions can be the normal range of environmental information, and once the detected environmental information exceeds the preset normal range, it is determined to be abnormal environmental information.

[0109] As an alternative example, environmental information acquisition: real-time acquisition of environmental information through built-in sensors of electronic devices, wherein the environmental information includes at least: light intensity, ambient temperature, and ambient humidity.

[0110] As an alternative example, the adjustment of feature weight combinations can be achieved through a dynamic weight adjustment mechanism, which specifically includes: introducing an adaptive weight adjustment algorithm into the dynamic weight adjustment mechanism, and automatically adjusting the feature weights corresponding to each biological static feature based on the environmental parameters collected in real time by the device sensors and the recognition confidence level fed back by the biometric recognition module (such as a feature recognition model used to identify biological static features).

[0111] Optionally, when the light intensity is <50 lux or >10000 lux (strong light / weak light environment), the recognition confidence of facial features is likely to be lower than 0.7. At this time, the feature weight corresponding to the facial features is automatically reduced (reduced by 0.1 each time, down to a minimum of 0.1), while the feature weights corresponding to fingerprint features and iris features are increased (e.g., the fingerprint weight is increased to a maximum of 0.5, and the iris weight is increased to a maximum of 0.4).

[0112] For example, in a strong outdoor light environment (light intensity 15000 lux), the device's sensors detect excessive light intensity, causing the confidence level for facial feature recognition to drop to 0.65. The adaptive algorithm automatically reduces the feature weight of the facial feature from 0.3 to 0.1, the feature weight of the fingerprint feature from 0.4 to 0.5, and the feature weight of the iris feature from 0.3 to 0.4. At this point, the confidence level for the fused biometric features can be maintained above 0.9.

[0113] Optionally, when the ambient humidity is greater than 70%RH (humid environment), the recognition confidence of fingerprint features is likely to be lower than 0.7. In this case, the feature weight of fingerprint features is automatically reduced (by 0.1 each time, down to a minimum of 0.2), while the feature weights of face features and iris features are increased (e.g., the face weight is increased to a maximum of 0.4, and the iris weight is increased to a maximum of 0.4).

[0114] For example, in a humid environment (85% RH), the recognition confidence of fingerprint features drops to 0.62. The algorithm reduces the feature weight of fingerprint features from 0.4 to 0.2, increases the feature weight of face features from 0.3 to 0.4, and increases the feature weight of iris features from 0.3 to 0.4. After fusion, the recognition confidence of biometric fusion features remains above 0.88.

[0115] As an optional embodiment, before querying the private key pre-allocated to the target object based on the bio-fusion features when the bio-dynamic features pass the liveness detection, the method further includes: detecting whether the bio-dynamic features of each modality meet the corresponding dynamic feature verification standard; and determining that the bio-dynamic features pass the liveness detection when the bio-dynamic features of each modality meet the corresponding dynamic feature verification standard.

[0116] The embodiments described above further improve the biometric liveness detection process, ensuring the authenticity and validity of biometric dynamic features. By detecting the biometric dynamic features of each modality and determining whether they meet their respective dynamic feature verification standards, the activity state of biometric dynamic features is detected, avoiding the misuse of inactive biometric features such as static images or videos. Only when the biometric dynamic features of fingerprints, faces, and irises all pass their respective dynamic feature verification standards—that is, changes in fingerprint pressure, subtle facial expressions, and subtle iris tremors—can the biometric dynamic features be confirmed to have passed liveness detection, thus allowing the subsequent private key query and signature generation process to continue. This further enhances the security of biometric collection, effectively prevents attacks that forge biometric features, lays the foundation for building a reliable electronic signature system, effectively enhances the accuracy of identity authentication and the unforgeability of electronic signatures, and also ensures the security and reliability of the electronic signature system, solving the technical problem of poor security in existing electronic signature technologies.

[0117] As an optional embodiment, when the biodynamic features of each modality meet the corresponding dynamic feature verification criteria, determining that the biodynamic features pass the liveness detection includes: detecting whether the time difference between the collection timestamp of the multimodal biofeature and the blockchain timestamp does not exceed a preset time threshold, wherein the blockchain timestamp is obtained by initiating a timestamp query request to the blockchain network, and the timestamp query request is initiated immediately after the collection of the biostatic features; if the time difference does not exceed the preset time threshold, the biodynamic features are determined to pass the liveness detection.

[0118] In the embodiments described above, to enhance the authenticity of biometric information and prevent replay attacks, a time synchronization verification mechanism is designed. Once the dynamic biometric features of fingerprints, faces, and irises all meet their respective dynamic feature verification standards, the mechanism further checks whether the time difference between the biometric data collection timestamp and the blockchain timestamp does not exceed a preset time threshold. This preset time threshold ensures that multimodal biometric features are collected within an acceptable time window. Specifically, the time difference between the collection timestamp generated immediately after the static biometric feature collection and the current blockchain time must be controlled within a certain range. By initiating a timestamp query request in real time and comparing the result with the collection timestamp, the timeliness of the multimodal biometric features can be determined, thereby deciding whether to pass the liveness detection. This precise time synchronization strategy not only improves the ability to identify fraudulent activities but also enhances the system's verification accuracy for biometric features, ensuring that only multimodal biometric features within the valid time frame can be used in the signing process. This significantly reduces potential security risks and improves the overall security and reliability of electronic signatures.

[0119] As an optional example, after the biometric acquisition device completes the acquisition of dynamic biometric features (or multimodal biometric features, or static biometric features), it immediately generates an acquisition timestamp and transmits the acquisition timestamp and biometric data (such as multimodal biometric features, static biometric features, or dynamic biometric features) synchronously to the blockchain interaction module through an encrypted channel. After receiving the data (such as multimodal biometric features, static biometric features, or dynamic biometric features), the blockchain interaction module initiates a timestamp query request to the blockchain network to obtain the current block timestamp of the blockchain network. It compares the acquisition timestamp with the blockchain timestamp to obtain a time verification result (such as the time difference between the acquisition timestamp and the blockchain timestamp). If the time verification result shows that the difference between the two is within 500ms, the time verification is considered successful. If the difference exceeds 500ms, an anomaly prompt is automatically triggered, requiring the target object to re-acquire biometric features to prevent replay attacks using historical biometric data.

[0120] The embodiments described above in this application, based on a comprehensive analysis of dynamic verification results (such as detecting whether the biometric dynamic features of each modality meet the corresponding dynamic feature verification standards) and time verification results (detecting whether the time difference between the collection timestamp of the multimodal biometric features and the blockchain timestamp does not exceed a preset time threshold), yield a dynamic biometric anti-tampering mechanism. Only when both dynamic and time verifications pass are the biometric features considered authentic, valid, and untampered. In the electronic signature system, this anti-tampering mechanism permeates the entire process of biometric feature collection, verification, signature generation, and verification, providing comprehensive protection for system security. This innovative anti-tampering mechanism effectively resists various forgery, tampering, and replay attacks, ensuring the security and credibility of the electronic signature system and providing solid technical support for the widespread application of electronic signatures.

[0121] As an optional embodiment, when the biodynamic features pass the liveness detection, allocating a private key to the target object based on the biofusion features includes: hashing the biofusion features to obtain a biometric hash value; verifying the biometric hash value against pre-set biometric constraints to obtain a feature matching value; determining a corresponding feature matching threshold based on a pre-set risk level for the signing information, wherein each risk level has a pre-set feature matching threshold; and allocating a private key to the target object when the feature matching value is greater than the feature matching threshold.

[0122] The embodiments described above in this application obtain biometric hash values ​​by hashing biometric fusion features, and then use these hash values ​​for risk level authentication and target object identification. Compared to directly using biometric fusion features for risk level authentication and target object identification, this avoids the tampering or theft of biometric fusion features. Furthermore, it fully utilizes the characteristics of hash algorithms, where even small changes in input data lead to significant differences in output hash values, effectively preventing unauthorized access and the misuse of biometric information. This not only improves the accuracy of biometric verification but also dynamically adjusts the matching threshold according to the risk level of different scenarios, enhancing the system's adaptability and security. It provides a reliable biometric authentication foundation for the implementation of electronic signatures, improves the security and reliability of electronic signatures, and solves the technical problem of poor security in existing electronic signature technologies.

[0123] Optionally, hashing the bio-fusion features includes: encrypting the fused bio-fusion features using the SHA-256 encryption algorithm to generate a unique bio-feature hash value.

[0124] Optionally, biometric constraints may include, but are not limited to, biometric types, quality standards, and dynamic features. A dynamic threshold mechanism is employed during the verification process, setting different matching thresholds based on different business risk levels. Higher security requirements correspond to higher risk levels, and consequently, higher matching values ​​are required between the biometric hash value and the biometric constraints.

[0125] As an optional example, when the target object initiates an application for an electronic signature, the collected unique biometric features (such as biometric fusion features) are matched and verified with preset biometric constraints to obtain a verification result; based on the verification result, the smart contract for the electronic signature is automatically triggered after the biometric verification is passed.

[0126] Optionally, the smart contract first verifies the target object's signing permission. By using pre-defined permission rules, it checks whether the target object has the authority to sign the information. If the permission verification passes, the smart contract will continue to execute the subsequent signing process. If the permission verification fails, the signing request will be rejected, and the target object will be notified of the corresponding prompts to ensure the legality and compliance of the signing operation.

[0127] Optionally, the smart contract invokes an encryption algorithm (such as ECC elliptic curve cryptography) to sign the hash value of the signing information (i.e., the hash value of the signing information) using the target object's private key, generating an electronic signature. Simultaneously, the hash value of the signing information, the electronic signature, the object's biometric digest (to protect biometric privacy, only the digest is uploaded to the blockchain), the signing timestamp, and other information are merged into a transaction module (or signature information). The constructed transaction module is then sent to the blockchain network for consensus verification.

[0128] As an optional embodiment, when the feature matching value is greater than the feature matching threshold, the private key allocated to the target object includes: determining identity information that matches the biometric fusion feature, wherein the identity information is pre-configured with identity matching conditions that match the biometric fusion feature; determining the target object's signing permission based on the identity information, wherein the identity information is also pre-configured with corresponding signing permissions; determining the permission verification conditions for the signing information in a pre-set permission rule, wherein the permission rule pre-sets multiple permission verification conditions and a level matching relationship between each permission verification condition and the signing information; and allocating the private key to the target object when the signing permission meets the permission verification conditions corresponding to the signing information.

[0129] In the embodiments described above, the biometric fusion feature is unique and can represent the identity information of the target object; the biometric hash value determined based on the biometric fusion feature is also unique and can also represent the identity information of the target object. Therefore, after the biometric hash value is verified by the feature matching value, the biometric hash value can be used as the identity information of the target object to further determine the signing authority of the target object. Then, if the signing authority meets the authority verification conditions corresponding to the signing information, a private key is allocated to the target object to ensure that the generation and verification process of the electronic signature follows the established authority rules, thus guaranteeing the legality and security of the electronic signature. Thus, through the dynamic authority verification and private key allocation mechanism, unauthorized signing behavior is effectively avoided, improving the compliance and user experience of the electronic signature system. At the same time, through the dynamic matching of authority, the flexibility and adaptability of the signing process are enhanced, ensuring the secure and effective application of the electronic signature in signing information at different levels, achieving the technical effect of improving the security of the electronic signature, and thus solving the technical problem of poor security of the electronic signature in the prior art.

[0130] Optionally, multiple permission verification conditions and the level matching relationship between each permission verification condition and the signing information are pre-set in the permission rules to realize dynamic matching and adjustment of permissions.

[0131] As an optional embodiment, encrypting the signing information with a private key to obtain an electronic signature includes: hashing the signing information to obtain a hash value of the signing information; and encrypting the hash value of the signing information with a private key to obtain an electronic signature.

[0132] The embodiments described above in this application obtain a hash value for the signed information by hashing the signed information, and then generate an electronic signature based on the hash value. Compared with the scheme of directly generating an electronic signature using the signed information, this avoids the leakage or tampering of the signed information. Furthermore, it fully utilizes the characteristics of hash algorithms, namely that even a small change in the input data will lead to a huge difference in the output hash value, thereby ensuring the integrity and immutability of the signed information, further ensuring the security and uniqueness of the electronic signature, achieving the technical effect of improving the security of the electronic signature, and thus solving the technical problem of poor security of electronic signatures in the prior art.

[0133] It should be noted that because the private key is held only by the target object and cannot be copied or cracked by other objects, this greatly enhances the anti-counterfeiting capability of electronic signatures.

[0134] In the embodiments described above, the generation of electronic signatures is not only fast, but also enhances the authenticity and non-repudiation of signatures while ensuring security, providing strong technical support for document signing in digital office and e-commerce.

[0135] It should be noted that in the subsequent electronic signature verification process, the electronic signature is decrypted using the public key to recover the hash value, which is then compared with the hash value of the current file. This can efficiently and accurately verify the validity of the signature, ensuring the integrity and reliability of the entire signing process.

[0136] As an optional embodiment, after encrypting the signing information using a private key to obtain an electronic signature, the method further includes: generating signature information based on the electronic signature, wherein the signature information includes at least the electronic signature; and uploading the signature information to a blockchain network.

[0137] In the above embodiments of this application, signature information is generated based on electronic signatures. This signature information includes at least an electronic signature. The signature information is then uploaded to a blockchain network. The distributed ledger technology of blockchain ensures the secure storage and immutability of the signature information, enabling transparent management of the electronic document signing process. This facilitates subsequent traceability and verification, improves the legal validity and security of electronic signatures, and achieves the technical effect of improving the security of electronic signatures. This solves the technical problem of poor security of electronic signatures in the prior art.

[0138] Optionally, and of course, in other embodiments, the content of the signature information can be further expanded, such as by adding the signer's identity information, signing timestamp, etc., to provide richer background information on the signature and enhance the completeness and credibility of the signature information.

[0139] It should be noted that signature information is data used to prove that the information has been signed by the target object, and typically includes: electronic signature, signer's identity information, signing timestamp, etc. The signature information may also contain the signer's public key certificate, biometric digest, and other data to provide richer background information, enhance the integrity and credibility of the signature information, and further verify the signer's identity.

[0140] The embodiments described above in this application, by uploading signature information to a blockchain network for storage, can provide comprehensive support for the verification and auditing of electronic signatures, ensuring the effectiveness and legal recognition of electronic signatures in various application scenarios. They also enhance the automation and intelligence of the signing process, making the entire electronic signature process more efficient and convenient.

[0141] Optionally, when the signature information is uploaded to the blockchain network, the nodes in the blockchain network will verify the transaction module (or signature information) to ensure its legality and validity. Once consensus is successful, the transaction module (or signature information) will be written into the blockchain block, completing the on-chain storage of the signature information. The on-chain signature information has the characteristics of being tamper-proof and traceable, providing strong protection for the legal validity and security of electronic signatures.

[0142] As an optional embodiment, after uploading the signature information to the blockchain network, the method further includes: hashing the signature information to be verified to obtain a verification hash value; extracting an electronic signature from the signature information on the blockchain network as the signature to be verified; decrypting the signature to be verified using a public key pre-configured for the target object to obtain a decryption hash value; and determining that the signature information to be verified has not been tampered with if the verification hash value and the decryption hash value are consistent.

[0143] In the above embodiments of this application, since the electronic signature is the result of encrypting the hash value of the signing information using the private key of the target object, when it is necessary to verify whether the original signing information has been tampered with, the signing information to be verified can be used as the information to be verified and processed using the same hash processing method to obtain the verification hash value. Since the signature information on the blockchain network is protected by the consensus verification of the blockchain network, the electronic signature can be extracted from the signature information on the blockchain network based on the signature information coordinate verification standard on the blockchain network, and decrypted using the public key of the target object to obtain the hash value of the original signing information. Therefore, the decryption hash value is the same as the hash value of the signing information, and the verification hash value is compared with the decryption hash value. If the two are consistent, it can be determined that the signing information to be verified has not been tampered with since the electronic signature was generated, thereby verifying the validity of the electronic signature and the original integrity of the signing information, ensuring the unforgeability of the electronic signature, and effectively verifying the untampered state of the signing information, providing a reliable verification basis for subsequent relevant parties, achieving the technical effect of improving the security of electronic signatures, and thus solving the technical problem of poor security of electronic signatures in the prior art.

[0144] This invention also provides an optional embodiment, which offers an electronic signature method. By integrating blockchain technology with multimodal biometric recognition technology, a decentralized electronic signature system is constructed to improve the security and reliability of electronic signatures, preventing signature forgery and identity misuse. A dynamic biometric collection and verification mechanism is designed to achieve real-time authentication of the object's identity and dynamic binding of the signing behavior, ensuring the authenticity and uniqueness of the signing behavior. The electronic signature process is automated, including document hash calculation, signature generation, and on-chain evidence storage, improving signing efficiency and optimizing the user experience. Cross-platform electronic signature standards and interface specifications are established to improve the cross-platform compatibility and interoperability of electronic signatures, promoting the widespread application of electronic signature technology. A complete evidence chain system is constructed, leveraging the immutability of blockchain and the uniqueness of multimodal biometrics to ensure the legal validity of electronic signatures in judicial practice and reduce dispute resolution costs.

[0145] Figure 2 This is an illustration of an electronic signature according to an embodiment of the present invention. Figure 1 ,like Figure 2As shown, it includes the following steps:

[0146] Step S201: Collect biometric features, obtain fingerprint, facial and iris biometric information and signature information.

[0147] Optionally, fingerprint, face, and iris biometric information are fused and encrypted to generate a unique biometric feature; then, a weighted fusion algorithm is used to fuse the preprocessed fingerprint, face, and iris features (where the feature weights are dynamically adjusted according to different scenario requirements, for example, in a financial scenario, the feature weight of the iris feature can be set to 0.4, the feature weight of the fingerprint feature can be set to 0.3, and the feature weight of the face feature can be set to 0.3) to generate biometric data (i.e., biometric fusion features).

[0148] Step S202: Generate a unique biometric hash.

[0149] Optionally, during the feature fusion process, an irreversible encryption algorithm (SHA-256) is used to encrypt the biometric fusion features to ensure the security of the template. The biometric data (i.e., the biometric fusion features) is converted into a fixed-length hash value to obtain a unique biometric feature (i.e., the biometric hash value). The unique biometric feature cannot be used to reconstruct the original biometric data from the hash value, thereby ensuring the security of the biometric template, preventing the biometric information from being leaked or tampered with during storage and transmission, and effectively protecting the biometric privacy of the object.

[0150] Step S203: The blockchain processes the signing information.

[0151] Optionally, the signing information is input into the blockchain model, which calculates the hash value of the signing information using the SHA-256 hash algorithm to generate a unique hash value (i.e., the signing information hash value). This hash value uniquely identifies the content of the signing information, and any modification to the signing information will result in a change in the hash value.

[0152] Step S204, Biometrics and Permission Verification.

[0153] Optionally, throughout the signing process, the smart contract verifies the identity and permissions of the target object according to preset permission rules to ensure the legality and compliance of the signing operation.

[0154] Optionally, various abnormal situations can be monitored in real time during the signing process.

[0155] For example, when the number of failed biometric verifications exceeds a preset threshold, the system automatically locks the target's account and sends an alert to the administrator; when the blockchain network malfunctions, the system attempts to reconnect to the network or switch to a backup node to ensure that the signature information is successfully uploaded to the blockchain for evidence storage.

[0156] It should be noted that the smart contract has a built-in permission rule library, which pre-sets multiple permission rules, including permission configurations for dimensions such as object role, type of document that can be signed, and signing permission level.

[0157] Optionally, the permission verification process is as follows: after the smart contract receives the signature application from the target object, it calls the object's identity information and biometric digest stored in the blockchain and matches them with the currently collected unique biometric hash value; after the match is successful, it determines whether the target object has the signature permission for the file (or signature information) based on the file type (such as the permission verification conditions for the signature information) and the target object's query permission rule base.

[0158] For example, ordinary employees only have the authority to sign labor contracts individually, while procurement contracts require the joint signature of the administrator and the finance personnel.

[0159] Optionally, if the permission verification passes, the smart contract generates a permission verification pass certificate and stores it on the blockchain; if the permission verification fails, the smart contract automatically generates a rejection reason, feeds it back to the object through the front-end interface, and records the rejection on the blockchain for evidence storage, which is convenient for subsequent auditing.

[0160] Step S205: Generate an electronic signature and upload it to the blockchain.

[0161] Optionally, the signing information of the document is input into the blockchain model, and the hash value of the signing information is calculated using the SHA-256 hash algorithm to generate a unique hash value.

[0162] Optionally, the target object initiates a signature request, and the electronic signature device matches and verifies the unique biometric features against preset biometric constraints. A dynamic threshold mechanism is used during the verification process, setting a matching threshold based on the business risk level of the document signing. Upon successful verification, the electronic signature smart contract is automatically triggered.

[0163] Optionally, the smart contract verifies the target object's signing authority. After confirming that the target object has the authority to sign the file (such as signing information), it calls the ECC elliptic curve cryptography algorithm to use the employee's target object to sign the hash value of the signing information, generating an electronic signature.

[0164] Optionally, the hash value of the signed information, the electronic signature, the object's biometric digest, the signing timestamp, and other information can be merged into a transaction module (i.e., signature information) and sent to the blockchain network for consensus verification. After successful consensus, the transaction module is written into the blockchain block, completing the on-chain storage of the signature information.

[0165] Step S206: Verify the electronic signature.

[0166] Optionally, the hash value of the signature information to be verified is recalculated to obtain the verification hash value, ensuring that the signature information is not tampered with during transmission and storage.

[0167] Optionally, based on the hash value verification result, the blockchain interaction module queries the blockchain containing the signed information to extract the hash value, signature information, and signature timestamp of the signed information to obtain an electronic signature; then, the public key of the target object is used to decrypt the electronic signature to obtain the original hash value of the signed information (i.e., the decrypted hash value), and then the decrypted hash value is compared with the recalculated verification hash value.

[0168] Step S207, hash value comparison.

[0169] Optionally, if the decryption hash value matches the verification hash value, it indicates that the signing information has not been tampered with during transmission and storage, and the electronic signature is valid. The verification result will be output. If the decryption hash value does not match the verification hash value, it indicates that the electronic signature has been tampered with or forged. The verification result will be output, and corresponding security measures will be taken, such as issuing an alarm and recording anomalies, to ensure the security and credibility of the electronic signature.

[0170] Step S208: If the hash value passes verification, perform dynamic anti-tampering and automated processing.

[0171] Optionally, a dynamic anti-tampering mechanism is used to perform liveness detection on the target object. For example, when collecting facial data, employees are required to perform actions such as blinking and opening their mouths to ensure that it is a real person operating the device; when collecting iris data, liveness detection is performed by detecting the pupil's reaction to light.

[0172] Specifically, this involves performing liveness detection on the unique multimodal biometric feature (or its dynamic features) during the acquisition process to ensure that the biometric features originate from a real person and are not forged. Specifically, dynamic features (such as changes in fingerprint pressure, subtle facial expressions, and minute tremors of the iris) are extracted and combined with static features for verification to obtain dynamic verification results. This process effectively increases the difficulty of forgery and prevents attacks using photos, videos, fingerprint films, etc., to forge biometric features. For example, facial liveness detection detects actions such as blinking and opening the mouth, while iris liveness detection detects the pupil's reaction to light.

[0173] Optionally, the dynamic anti-tampering mechanism can also be implemented by matching the biometric collection time with the blockchain timestamp. If the time difference between the biometric collection time and the blockchain timestamp does not exceed a preset time threshold, it means that the dynamic anti-tampering mechanism has passed time verification, ensuring the timeliness of biometrics and preventing replay attacks using historical biometric data.

[0174] Optionally, a biometric is considered authentic, valid, and untampered only if both dynamic and time-based verifications pass.

[0175] Optionally, automated processing is achieved through smart contracts. Throughout the signing process, the smart contract verifies the employee's identity and permissions according to preset permission rules, ensuring the legality and compliance of the signing operation. Specifically, based on a biometric dynamic feature anti-tampering mechanism, the smart contract verifies the target object's identity and permissions according to preset permission rules. By accessing the target object's identity information and biometric data, combined with rules from the permission rule base, the smart contract can quickly and accurately determine whether the target object has the permission to sign the relevant information. If the target object's permission verification passes, the smart contract continues the signing process; if the permission verification fails, the smart contract automatically rejects the signing request and provides the target object with detailed reasons for rejection, such as "the object does not have permission to sign this document," thereby ensuring the legality and compliance of the signing operation and preventing unauthorized signing.

[0176] As an optional implementation, the smart contract has a built-in anomaly monitoring module to monitor various anomalies during the signing process in real time. The monitoring dimensions and judgment criteria are as follows:

[0177] Biometric verification anomaly: Set a verification failure threshold (e.g., 3 consecutive verification failures). If an object triggers the threshold consecutively within 10 minutes, the monitoring module determines it as "biometric verification anomaly" and records the reason for each verification failure.

[0178] Blockchain network anomaly: Monitor the block generation time (normal range is 10s-30s) and node response latency (normal range is within 500ms) of the blockchain network. If the block generation time exceeds 60s or the node response latency exceeds 1s, it is determined to be "blockchain network anomaly".

[0179] Signature process timeout exception: Set a timeout period for the signature process (e.g., 30 minutes). If the entire process from the object initiating the signature request to the transaction module's on-chain notarization exceeds the timeout period, it will be judged as "signature process timeout exception".

[0180] As an optional implementation, the smart contract automatically triggers corresponding exception handling procedures for different types of abnormal situations. The specific measures are as follows:

[0181] For biometric verification anomalies: The target account is automatically locked, and its signing privileges are suspended for 1 hour (first anomaly), 2 hours (second anomaly), and 24 hours (third or more anomalies). An alert is sent to the system administrator, and an anomaly operation log is generated and stored on the blockchain. During the lockout period, the target can initiate an identity reset request, which requires manual review by the administrator (combining the target's ID card information and historical biometric data) before the account can be unlocked.

[0182] In response to blockchain network anomalies: The smart contract automatically attempts to reconnect to the blockchain network, with a 5-second interval between each reconnection, and a maximum of 5 attempts. If a reconnection fails, it switches to a backup blockchain node (at least 3 backup nodes are preset, distributed in different regions). After switching nodes, the consensus verification request for the transaction module is re-initiated, and the node switching process is recorded. If all backup nodes fail to connect, the smart contract pauses the signing process, sends a "Network error, please try again later" message to the target, and automatically resumes the signing process once the network is restored.

[0183] In response to timeout anomalies in the signing process: the smart contract automatically terminates the current signing process, releases system resources, and sends a timeout reminder to the recipient; the recipient can choose to re-initiate the signing application, and the system automatically retrieves historical biometric data to reduce repetitive operations. Timeout records are stored on the blockchain, and the smart contract periodically compiles timeout anomaly data, analyzes the causes of timeouts, and pushes optimization suggestions to the administrator.

[0184] As an optional example, the system monitors various anomalies during the signing process in real time. For instance, if the number of failed biometric verifications exceeds a preset threshold, indicating a malicious attack or abnormal object operation, the employee's account is automatically locked, and an alert is sent to the administrator. Alternatively, if the blockchain network experiences anomalies, such as excessive network latency or abnormal block generation time, affecting the normal on-chain storage of signed information, the system attempts to reconnect to the network or switches to a backup node to ensure successful on-chain storage of signed information. The smart contract uses a built-in monitoring module to detect and record these anomalies in real time.

[0185] Optionally, once an anomaly is detected, the corresponding anomaly handling process is automatically triggered. In cases of excessive biometric verification failures, the account of the affected individual is automatically locked, their signing permissions are suspended, and an alert is sent to the system administrator. Simultaneously, a detailed anomaly operation log is recorded for subsequent security audits and investigations. In the event of blockchain network anomalies, measures such as attempting to reconnect to the network, waiting for network recovery, or switching to a backup blockchain node are implemented to ensure that the signed information can be successfully uploaded to the blockchain for notarization. This automated anomaly handling mechanism enables timely detection and response to various anomalies during the signing process, ensuring stable system operation and secure signing operations, thereby improving system reliability and customer experience.

[0186] The embodiments described above in this application primarily generate unique biometric features by collecting and encrypting multimodal biometric data (fingerprints, faces, and irises), and then combining this with blockchain technology to generate, verify, and store electronic signatures. The core principle is to ensure the security and immutability of the signature through multimodal biometric recognition and blockchain technology. By integrating blockchain and multimodal biometric recognition technologies, a decentralized electronic signature system is constructed, achieving highly secure signatures and preventing identity theft and forgery. This system is suitable for digital office and e-commerce scenarios, aiming to improve the security, reliability, and legal validity of electronic signatures.

[0187] Figure 3 This is an illustration of an electronic signature according to an embodiment of the present invention. Figure 2 ,like Figure 3 As shown, multimodal biometrics include at least fingerprint features, facial features, and iris features. The process of generating an electronic signature based on fingerprint features, facial features, and iris features includes the following steps:

[0188] Step S311: Iris acquisition, obtaining iris features;

[0189] Step S321: Face capture, obtaining facial features;

[0190] Step S331: Fingerprint acquisition, obtaining fingerprint features;

[0191] Step S302: The image quality of the collected iris features, face features and fingerprint features is judged respectively. If the judgment result is unqualified, the image is collected again; if the judgment result is qualified, the process proceeds to step S303.

[0192] Step S303, biometric fusion and encryption, that is, weighted fusion of iris features, facial features and fingerprint features to obtain biometric fusion features.

[0193] Step S304: Generate a unique biometric hash value, that is, perform hash processing on the biometric fusion feature.

[0194] Step S305: Generate an electronic signature. This is achieved by using a biometric hash value to obtain a private key, and then using the private key to encrypt the hashed signature information (i.e., the signature information hash value) to obtain an electronic signature.

[0195] According to an embodiment of the present invention, an embodiment of an electronic signature device is also provided. It should be noted that the electronic signature device can be used to execute the electronic signature method in the embodiment of the present invention, and the electronic signature method in the embodiment of the present invention can be executed in the electronic signature device.

[0196] Figure 4 This is a schematic diagram of an electronic signature device according to an embodiment of the present invention, such as... Figure 4 As shown, the device may include: a collection module 40 for collecting multimodal biometrics of a target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; a fusion module 42 for fusing the multiple modalities of static biometrics into a biometric fusion feature; an allocation module 44 for allocating a private key to the target object based on the biometric fusion feature when the dynamic biometrics pass liveness detection; and an encryption module 46 for encrypting the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0197] The embodiments described above in this application collect multimodal biometric features of the target object and use a dynamic weight adjustment algorithm to fuse multimodal static features into biometric fusion features. This ensures high-precision biometric recognition even in complex environments. Furthermore, when the biometric dynamic features pass liveness detection, a private key is allocated to the target object based on the biometric fusion features. This ensures that the allocation of the private key is strictly based on real-time and authentic biometric features, greatly improving the security of the private key. It also ensures that the electronic signature encrypted with this private key has high personal attributes and is not easily copied, effectively preventing the forgery of electronic signatures and the impersonation of the target object's identity. This ensures the authenticity and uniqueness of the signing behavior, thereby achieving the technical effect of improving the security of electronic signatures and solving the technical problem of poor security of electronic signatures in the prior art.

[0198] It should be noted that the acquisition module 40 in this embodiment can be used to execute step S102 in this application embodiment, the fusion module 42 in this embodiment can be used to execute step S104 in this application embodiment, the allocation module 44 in this embodiment can be used to execute step S106 in this application embodiment, and the encryption module 46 in this embodiment can be used to execute step S108 in this application embodiment. The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments.

[0199] The embodiments described above in this application collect multimodal biometric features of the target object and use a dynamic weight adjustment algorithm to fuse multimodal static features into biometric fusion features. This ensures high-precision biometric recognition even in complex environments. Furthermore, when the biometric dynamic features pass liveness detection, a private key is allocated to the target object based on the biometric fusion features. This ensures that the allocation of the private key is strictly based on real-time and authentic biometric features, greatly improving the security of the private key. It also ensures that the electronic signature encrypted with this private key has high personal attributes and is not easily copied, effectively preventing the forgery of electronic signatures and the impersonation of the target object's identity. This ensures the authenticity and uniqueness of the signing behavior, thereby achieving the technical effect of improving the security of electronic signatures and solving the technical problem of poor security of electronic signatures in the prior art.

[0200] As an optional embodiment, the fusion module includes: an identification unit for identifying business scenarios for collecting multimodal biometrics, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to multiple modal biometric static features respectively; and a fusion unit for weighted fusion of multiple modal biometric static features according to the feature weight combination corresponding to the business scenario to obtain biometric fusion features.

[0201] As an optional embodiment, the apparatus further includes: a first determining unit, configured to determine the recognition confidence of each modality's biostatic features before weighted fusion of multiple modalities' biostatic features according to the feature weight combination corresponding to the business scenario to obtain biofused features, wherein the recognition confidence is used to represent the degree of matching between the biostatic features and template features in a preset feature library; a second determining unit, configured to determine biostatic features with recognition confidence lower than a first recognition threshold as features to be adjusted; and an adjusting unit, configured to reduce the corresponding feature weights of the features to be adjusted and increase the corresponding feature weights of biostatic features that do not belong to the features to be adjusted, thereby obtaining a feature weight combination.

[0202] As an optional embodiment, the adjustment unit includes: a statistics subunit, used to count the number of consecutive historical instances where the recognition confidence of the feature to be adjusted is continuously lower than a first recognition threshold; a first adjustment subunit, used to reduce the feature weight corresponding to the feature to be adjusted to a preset minimum weight when the number of consecutive historical instances is greater than a preset number threshold; a second adjustment subunit, used to increase the corresponding feature weight for biostatic features with a recognition confidence greater than a second confidence threshold; and a generation subunit, used to generate a feature weight combination based at least on the preset minimum weight corresponding to the feature to be adjusted and the increased feature weight.

[0203] As an optional embodiment, the device further includes: a first detection unit, used to detect and collect various environmental information of multimodal biofeatures before reducing the corresponding feature weight for the feature to be adjusted and increasing the corresponding feature weight for biostatic features that do not belong to the feature to be adjusted, and obtaining the feature weight combination; a third determination unit, used to determine environmental information that does not meet the preset environmental conditions as abnormal environmental information; and a fourth determination unit, used to determine biostatic features whose recognition accuracy is negatively affected by abnormal environmental information as features to be adjusted.

[0204] As an optional embodiment, the device further includes: a first detection submodule, configured to detect whether the biodynamic features of each modality meet the corresponding dynamic feature verification standard before querying the private key pre-allocated to the target object based on the biofusion features when the biodynamic features pass the liveness detection; and a first determination submodule, configured to determine that the biodynamic features pass the liveness detection when the biodynamic features of each modality meet the corresponding dynamic feature verification standard.

[0205] As an optional embodiment, the first determining submodule includes: a second detection submodule, used to detect whether the time difference between the collection timestamp of the multimodal biometrics and the blockchain timestamp does not exceed a preset time threshold, wherein the blockchain timestamp is obtained by initiating a timestamp query request to the blockchain network, and the timestamp query request is initiated immediately after the collection of the biometric static features; the second determining submodule is used to determine that the biometric dynamic features pass the liveness detection if the time difference does not exceed the preset time threshold.

[0206] As an optional embodiment, the allocation module includes: a first hash unit, used to hash the bio-fusion feature when the bio-dynamic feature passes liveness detection, to obtain a bio-feature hash value; a second detection unit, used to detect the bio-feature hash value and verify it against pre-set bio-feature constraints, to obtain a feature matching value; a fifth determination unit, used to determine the corresponding feature matching threshold based on the risk level pre-set for the signing information, wherein each risk level has a pre-set corresponding feature matching threshold; and an allocation unit, used to allocate a private key to the target object when the feature matching value is greater than the feature matching threshold.

[0207] As an optional embodiment, the allocation unit includes: a first determining subunit, configured to determine identity information matching the biometric fusion feature when the feature matching value is greater than the feature matching threshold, wherein the identity information is pre-configured with identity matching conditions matching the biometric fusion feature; a second determining subunit, configured to determine the signing permission of the target object based on the identity information, wherein the identity information is also pre-configured with corresponding signing permissions; a third determining subunit, configured to determine the permission verification conditions for the signing information in a pre-set permission rule, wherein the permission rule pre-sets multiple permission verification conditions and a level matching relationship between each permission verification condition and the signing information; and an allocation subunit, configured to allocate a private key to the target object when the signing permission meets the permission verification conditions corresponding to the signing information.

[0208] As an optional embodiment, the encryption module includes: a second hash unit for hashing the signing information to obtain a hash value of the signing information; and an encryption unit for encrypting the hash value of the signing information using a private key to obtain an electronic signature.

[0209] As an optional embodiment, the device further includes: a generation submodule, used to encrypt the signing information using a private key to obtain an electronic signature, and then generate signature information based on the electronic signature, wherein the signature information includes at least an electronic signature; and an up-to-chain submodule, used to upload the signature information to a blockchain network.

[0210] As an optional embodiment, the device further includes: a hash submodule, used to perform hash processing on the signature information to be verified after uploading the signature information to the blockchain network to obtain a verification hash value; an extraction submodule, used to extract the electronic signature from the signature information on the blockchain network as the signature to be verified; a decryption submodule, used to decrypt the signature to be verified using a public key pre-configured for the target object to obtain a decryption hash value; and a third determination submodule, used to determine that the signature information to be verified has not been tampered with if the verification hash value and the decryption hash value are consistent.

[0211] Embodiments of the present invention can provide an electronic device, which can be a computer terminal, and the computer terminal can be any one of a group of computer terminal devices. Optionally, in this embodiment, the computer terminal can also be replaced by a mobile terminal or other terminal device.

[0212] Optionally, in this embodiment, the computer terminal may be located in at least one of a plurality of network devices in a computer network.

[0213] In this embodiment, the computer terminal described above can execute the program code for the following steps in the electronic signature method: collecting multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; fusing the multiple modalities of static biometrics into a biometric fusion feature; when the dynamic biometrics pass liveness detection, allocating a private key to the target object based on the biometric fusion feature; encrypting the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0214] Figure 5 This is a structural block diagram of a computer terminal according to an embodiment of the present invention, such as... Figure 5 As shown, the computer terminal 50 may include one or more (only one is shown in the figure) processors 52 and memory 54.

[0215] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the electronic signature method and apparatus in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned electronic signature method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal 50 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0216] The processor can invoke information and applications stored in the memory via a transmission device to perform the following steps: collecting multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; fusing the multiple modalities of static biometrics into a biofusion feature; if the dynamic biometrics pass liveness detection, allocating a private key to the target object based on the biofusion feature; encrypting the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0217] Optionally, the processor may also execute program code for the following steps: identifying business scenarios for collecting multimodal biometrics, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to multiple modal biometric static features respectively; and weighting and fusing the multiple modal biometric static features according to the feature weight combination corresponding to the business scenario to obtain biometric fusion features.

[0218] Optionally, the processor may also execute program code for the following steps: determining the recognition confidence of the biostatic features of each modality, wherein the recognition confidence is used to represent the degree of matching between the biostatic features and template features in a preset feature library; identifying biostatic features with recognition confidence lower than a first recognition threshold as features to be adjusted; reducing the corresponding feature weights for the features to be adjusted, and increasing the corresponding feature weights for biostatic features that do not belong to the features to be adjusted, thereby obtaining a feature weight combination.

[0219] Optionally, the processor may also execute program code for the following steps: counting the number of consecutive historical occurrences of the recognition confidence of the feature to be adjusted being lower than the first recognition threshold; when the number of consecutive historical occurrences is greater than a preset number threshold, reducing the feature weight corresponding to the feature to be adjusted to a preset minimum weight; increasing the corresponding feature weight for biostatic features with a recognition confidence greater than the second confidence threshold; and generating a feature weight combination based at least on the preset minimum weight corresponding to the feature to be adjusted and the increased feature weight.

[0220] Optionally, the processor may also execute program code that performs the following steps: detects and collects various environmental information of multimodal biometrics; identifies environmental information that does not meet preset environmental conditions as abnormal environmental information; and identifies biostatic features whose recognition accuracy is negatively affected by abnormal environmental information as features to be adjusted.

[0221] Optionally, the processor may also execute program code that performs the following steps: detects whether the biological dynamic features of each modality meet the corresponding dynamic feature verification criteria; and determines that the biological dynamic features pass the liveness detection if the biological dynamic features of each modality meet the corresponding dynamic feature verification criteria.

[0222] Optionally, the processor may also execute program code for the following steps: detecting whether the time difference between the collection timestamp of the multimodal biometric feature and the blockchain timestamp does not exceed a preset time threshold, wherein the blockchain timestamp is obtained by initiating a timestamp query request to the blockchain network, and the timestamp query request is initiated immediately after the collection of the static biometric feature; if the time difference does not exceed the preset time threshold, determining that the dynamic biometric feature has passed the liveness detection.

[0223] Optionally, the processor may also execute program code for the following steps: when the biodynamic features pass liveness detection, hash the biofusion features to obtain a biofeedback hash value; detect the biofeedback hash value and verify it against pre-set biofeedback constraints to obtain a feature matching value; determine the corresponding feature matching threshold based on the risk level pre-set for the signing information, wherein each risk level has a pre-set corresponding feature matching threshold; and allocate a private key to the target object when the feature matching value is greater than the feature matching threshold.

[0224] Optionally, the processor may also execute program code for the following steps: if the feature matching value is greater than the feature matching threshold, determine identity information that matches the biometric fusion feature, wherein the identity information is pre-configured with identity matching conditions that match the biometric fusion feature; determine the signing permission of the target object based on the identity information, wherein the identity information is also pre-configured with corresponding signing permissions; determine the permission verification conditions for the signing information in the pre-set permission rules, wherein the permission rules pre-set multiple permission verification conditions and the level matching relationship between each permission verification condition and the signing information; and allocate a private key to the target object if the signing permission meets the permission verification conditions corresponding to the signing information.

[0225] Optionally, the processor may also execute program code that performs the following steps: hashing the signing information to obtain a hash value; encrypting the hash value using a private key to obtain an electronic signature.

[0226] Optionally, the processor may also execute program code that performs the following steps: generating signature information based on the electronic signature, wherein the signature information includes at least: an electronic signature; and uploading the signature information to the blockchain network.

[0227] Optionally, the processor may also execute program code for the following steps: hashing the signature information to be verified to obtain a verification hash value; extracting an electronic signature from the signature information on the blockchain network as the signature to be verified; decrypting the signature to be verified using a public key pre-configured for the target object to obtain a decryption hash value; and determining that the signature information to be verified has not been tampered with if the verification hash value and the decryption hash value are consistent.

[0228] Those skilled in the art will understand that Figure 5 The structure shown is for illustrative purposes only. The computer terminal can also be a smartphone (such as an Android phone, an iOS phone, etc.), a tablet computer, a mobile internet device (MID), a PAD, and other terminal devices. Figure 5This does not limit the structure of the aforementioned electronic device. For example, computer terminal 50 may also include components that are more advanced than those described above. Figure 5 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 5 The different configurations shown.

[0229] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a computer program instructing the hardware related to the terminal device. The computer program can be stored in a non-volatile medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0230] Embodiments of the present invention also provide a non-volatile storage medium. Optionally, in this embodiment, the aforementioned non-volatile storage medium can be used to store the program code executed by the electronic signature method provided in the above embodiments.

[0231] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0232] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: collecting multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; fusing the multiple modalities of static biometrics into a bio-fusion feature; in the case that the dynamic biometrics pass liveness detection, allocating a private key to the target object based on the bio-fusion feature; encrypting the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

[0233] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: identifying business scenarios for collecting multimodal biometrics, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to multiple modal biometric static features respectively; and weighting and fusing the multiple modal biometric static features according to the feature weight combination corresponding to the business scenario to obtain biometric fusion features.

[0234] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: determining the recognition confidence of the biostatic features of each modality, wherein the recognition confidence is used to represent the degree of matching between the biostatic features and template features in a preset feature library; identifying biostatic features with recognition confidence lower than a first recognition threshold as features to be adjusted; reducing the corresponding feature weight for the features to be adjusted, and increasing the corresponding feature weight for biostatic features that do not belong to the features to be adjusted, thereby obtaining a feature weight combination.

[0235] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: counting the number of consecutive historical occurrences of the recognition confidence of the feature to be adjusted being lower than a first recognition threshold; when the number of consecutive historical occurrences is greater than a preset number threshold, reducing the feature weight corresponding to the feature to be adjusted to a preset minimum weight; increasing the corresponding feature weight for biostatic features with a recognition confidence greater than a second confidence threshold; and generating a feature weight combination based at least on the preset minimum weight corresponding to the feature to be adjusted and the increased feature weight.

[0236] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: detecting and collecting various environmental information of multimodal biometrics; identifying environmental information that does not meet preset environmental conditions as abnormal environmental information; and identifying biostatic features whose identification accuracy is negatively affected by abnormal environmental information as features to be adjusted.

[0237] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: detecting whether the biological dynamic features of each modality meet the corresponding dynamic feature verification criteria; and determining that the biological dynamic features pass the liveness detection if the biological dynamic features of each modality meet the corresponding dynamic feature verification criteria.

[0238] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: detecting whether the time difference between the collection timestamp of the multimodal biometric feature and the blockchain timestamp does not exceed a preset time threshold, wherein the blockchain timestamp is obtained by initiating a timestamp query request to the blockchain network, and the timestamp query request is initiated immediately after the collection of the biometric static feature; if the time difference does not exceed the preset time threshold, determining that the biometric dynamic feature has passed the liveness detection.

[0239] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: when the biodynamic features pass the liveness detection, the biofusion features are hashed to obtain a biofeedback hash value; the biofeedback hash value is matched and verified against pre-set biofeedback constraints to obtain a feature matching value; a corresponding feature matching threshold is determined based on the risk level pre-set for the signing information, wherein each risk level has a pre-set corresponding feature matching threshold; and a private key is allocated to the target object when the feature matching value is greater than the feature matching threshold.

[0240] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: when the feature matching value is greater than the feature matching threshold, determining identity information that matches the biometric fusion feature, wherein the identity information is pre-configured with identity matching conditions that match the biometric fusion feature; determining the signing permission of the target object based on the identity information, wherein the identity information is also pre-configured with corresponding signing permissions; determining the permission verification conditions for the signing information in the pre-set permission rules, wherein the permission rules pre-set multiple permission verification conditions and the level matching relationship between each permission verification condition and the signing information; and allocating a private key to the target object when the signing permission meets the permission verification conditions corresponding to the signing information.

[0241] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: hashing the signing information to obtain a hash value of the signing information; encrypting the hash value of the signing information using a private key to obtain an electronic signature.

[0242] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: generating signature information based on the electronic signature, wherein the signature information includes at least: an electronic signature; and uploading the signature information to the blockchain network.

[0243] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: hashing the signature information to be verified to obtain a verification hash value; extracting an electronic signature from the signature information on the blockchain network as a signature to be verified; decrypting the signature to be verified using a public key pre-configured for the target object to obtain a decryption hash value; and determining that the signature information to be verified has not been tampered with if the verification hash value and the decryption hash value are consistent.

[0244] Embodiments of the present invention also provide a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it implements the steps of the electronic signature method provided in the above embodiments.

[0245] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0246] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0247] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0248] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0249] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0250] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned non-volatile storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0251] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An electronic signature method, characterized in that, include: Collect multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; The biostatic features from multiple modalities are fused into a biofusion feature; If the biological dynamic characteristics pass the liveness detection, a private key is allocated to the target object based on the biological fusion characteristics; The signing information is encrypted using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

2. The method according to claim 1, characterized in that, Fusing the biostatic features from multiple modalities into a biofusion feature includes: Identify the business scenarios for collecting the multimodal biometrics, wherein different business scenarios correspond to different feature weight combinations, and each feature weight combination includes: feature weights corresponding to the biometric static features of multiple modalities respectively; Based on the feature weight combination corresponding to the business scenario, the biostatic features of multiple modalities are weighted and fused to obtain the biofusion feature.

3. The method according to claim 2, characterized in that, Before weighted fusing the biostatic features of multiple modalities according to the feature weight combination corresponding to the business scenario to obtain the biofused features, the method further includes: Determine the recognition confidence level of the biostatic feature for each modality, wherein the recognition confidence level is used to represent the degree of matching between the biostatic feature and template features in a preset feature library; The biostatic features whose recognition confidence is lower than the first recognition threshold are identified as features to be adjusted. The feature weights are decreased for the feature to be adjusted, and the feature weights are increased for the biological static features that do not belong to the feature to be adjusted, to obtain the feature weight combination.

4. The method according to claim 3, characterized in that, Decrease the corresponding feature weight for the feature to be adjusted, and increase the corresponding feature weight for the biostatic features that do not belong to the feature to be adjusted, to obtain the feature weight combination including: The number of consecutive historical instances in which the recognition confidence of the feature to be adjusted is continuously lower than the first recognition threshold is counted. If the number of consecutive historical occurrences exceeds a preset threshold, the feature weight corresponding to the feature to be adjusted will be reduced to a preset minimum weight. The corresponding feature weights are added to the biostatic features whose identification confidence is greater than the second confidence threshold; The feature weight combination is generated based at least on the preset minimum weight corresponding to the feature to be adjusted and the increased feature weight.

5. The method according to claim 3, characterized in that, Before decreasing the corresponding feature weight for the feature to be adjusted and increasing the corresponding feature weight for the biostatic features that do not belong to the feature to be adjusted, to obtain the feature weight combination, the method further includes: The system detects and collects various environmental information related to the multimodal biological characteristics. Environmental information that does not meet the preset environmental conditions is identified as abnormal environmental information; The biostatic features whose accuracy is negatively affected by the abnormal environmental information are identified as the features to be adjusted.

6. The method according to claim 1, characterized in that, Before querying the private key pre-assigned to the target object based on the bio-fusion characteristics, in the case where the bio-dynamic features pass the liveness detection, the method further includes: Detect whether the biological dynamic characteristics of each modality meet the corresponding dynamic characteristic verification criteria; If the biological dynamic features of each modality meet the corresponding dynamic feature verification criteria, the biological dynamic features are determined to pass the liveness detection.

7. The method according to claim 6, characterized in that, Determining that the biological dynamic features pass the liveness detection, provided that the biological dynamic features in each modality meet the corresponding dynamic feature verification criteria, includes: The time difference between the collection timestamp of the multimodal biometrics and the blockchain timestamp is detected to be no more than a preset time threshold. The blockchain timestamp is obtained by sending a timestamp query request to the blockchain network. The timestamp query request is sent immediately after the collection of the biometric static features is completed. If the time difference does not exceed the preset time threshold, the biological dynamic characteristics are determined to pass the liveness detection.

8. The method according to claim 1, characterized in that, When the biodynamic features pass the liveness detection, allocating a private key to the target object based on the biofusion features includes: When the biological dynamic features pass the liveness detection, the biological fusion features are hashed to obtain the biological feature hash value; The biometric hash value is matched and verified against pre-set biometric constraints to obtain a feature matching value; Based on the risk level preset for the signing information, a corresponding feature matching threshold is determined, wherein each risk level has a corresponding feature matching threshold preset. A private key is assigned to the target object when the feature matching value is greater than the feature matching threshold.

9. The method according to claim 8, characterized in that, When the feature matching value is greater than the feature matching threshold, the private key allocated to the target object includes: If the feature matching value is greater than the feature matching threshold, identity information matching the biometric fusion feature is determined, wherein the identity information is pre-configured with identity matching conditions matching the biometric fusion feature; Based on the identity information, the signature permissions of the target object are determined, wherein the identity information is also pre-configured with corresponding signature permissions; In the pre-set permission rules, the permission verification conditions for the signing information are determined, wherein the permission rules pre-set multiple permission verification conditions and the level matching relationship between each permission verification condition and the signing information; The private key is assigned to the target object if the signing permission meets the permission verification conditions corresponding to the signing information.

10. The method according to claim 1, characterized in that, The signature information is encrypted using the private key to obtain an electronic signature, including: The signature information is hashed to obtain the signature information hash value; The electronic signature is obtained by encrypting the hash value of the signing information using the private key.

11. The method according to claim 1, characterized in that, After encrypting the signing information using the private key to obtain an electronic signature, the method further includes: Based on the electronic signature, signature information is generated, wherein the signature information includes at least the electronic signature; The signature information is uploaded to the blockchain network.

12. The method according to claim 11, characterized in that, After uploading the signature information to the blockchain network, the method further includes: The signature information to be verified is hashed to obtain the verification hash value; Extract the electronic signature from the signature information on the blockchain network as a signature to be verified; The signature to be verified is decrypted using the public key pre-configured for the target object to obtain the decryption hash value; If the verification hash value matches the decryption hash value, it is determined that the signature information to be verified has not been tampered with.

13. An electronic signature device, characterized in that, include: The acquisition module is used to acquire multimodal biometrics of the target object, wherein the multimodal biometrics include at least: multiple modalities of static biometrics and multiple modalities of dynamic biometrics; A fusion module is used to fuse the biostatic features from multiple modalities into a biofusion feature; The allocation module is used to allocate a private key to the target object based on the bio-fusion characteristics when the bio-dynamic characteristics pass the liveness detection. An encryption module is used to encrypt the signing information using the private key to obtain an electronic signature, wherein the signing information is the original document to be signed.

14. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the electronic signature method of any one of claims 1 to 12 through the computer program.

15. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the electronic signature method according to any one of claims 1 to 12.

Images