Signature verification system and method based on image analysis

Abstract

A signature verification method based on image analysis, includes capturing handwriting data of a signatory during a signature process using a signature capture device, capturing image data of the signatory during the signing process using an image capturing device; obtaining at least one feature information from the image data; comparing the handwriting data and the at least one feature information with at least one authentication reference data to generate a comparison result; and determining authenticity of the signing process performed by the signatory based on the comparison result.

Description

[Technical Field]

[0001] This invention relates to a signature verification system and method based on image analysis, and particularly to a signature verification system and method that uses an image capturing device to capture image data during the signing process and thereby determine the authenticity of the signature. [Previous Technology]

[0002] Traditional signature verification methods mainly rely on static signature comparison. For example, when signing a document, the verifier usually needs to visually compare the current signature with a pre-stored signature template. This comparison method relies purely on the verifier's experience and judgment, which is prone to subjective errors. Moreover, static signature comparison cannot reflect the dynamic characteristics of the signing process, such as the speed and force of the signing, and is therefore easily forged.

[0003] With the development of technology, the application of electronic signature pads is becoming increasingly widespread. However, electronic signature pads can only record signature information on a two-dimensional plane and still lack verification of the signature process characteristics, making it difficult to effectively prevent imitation by others. As for the development of biometric identification technologies (such as fingerprint recognition, facial recognition, iris recognition, etc.), although they provide new solutions for identity verification, they usually require additional or special hardware support and cannot be effectively integrated with the signature verification process.

[0004] In view of this, how to provide a more reliable signature verification mechanism without the need for special hardware has become one of the goals that the industry is striving for. [Summary of the Invention]

[0005] Therefore, the present invention mainly provides a signature verification system and a signature verification method based on image analysis, so as to provide a more reliable signature verification mechanism.

[0006] This invention discloses a signature verification method based on image analysis, comprising: using a signature capturing device to capture handwriting data of a signer during a signature process; using an image capturing device to capture image data of the signer during the signature process; obtaining at least one feature information from the image data; comparing the handwriting data and the at least one feature information with at least one authentication reference data to generate a comparison result; and determining the authenticity of the signer's signature process based on the comparison result.

[0007] This embodiment of the invention further discloses a signature verification system based on image analysis, comprising a signature capturing device for capturing handwriting; an image capturing device for capturing images; a processing unit coupled to the signature capturing device and the image capturing device; and a storage unit coupled to the processing unit and storing a program code that instructs the processing unit to execute a signature verification method. The signature verification method includes the following steps: using the signature capturing device to capture handwriting data of a signer during a signing process; using the image capturing device to capture image data of the signer during the signing process; obtaining at least one feature information from the image data; comparing the handwriting data and the at least one feature information with at least one authentication reference data to generate a comparison result; and determining the authenticity of the signer's signing process based on the comparison result.

Implementation Method

[0008] Please refer to Figure 1, which is a schematic diagram of a signature verification system 10 according to an embodiment of the present invention. The signature verification system 10 verifies the authenticity of a signer's signature based on image analysis. It includes a signature capture device 101, an image capture device 100, a processing unit 102, and a storage unit 104. The image capture device 100 can be a camera or any device with image capture function, used to capture images. The image capture device 100 should be positioned in an appropriate location to clearly capture the signer's hand movements and the signing process. The signature capture device 101 can be an electronic signature pad, a touch panel, a pressure sensing plate, or other device capable of detecting or capturing handwriting. The processing unit 102 is coupled to the image capture device 100 and the signature capture device 101, and is used to receive and process image data from the image capture device 100 and handwriting data from the signature capture device 101. It can be a microprocessor, a digital signal processor, or other processor with computing power. Storage unit 104 is coupled to processing unit 102 and stores code 106 to instruct processing unit 102 to execute a signature verification process 20 to determine the authenticity of the signer's signature through image capture and analysis.

[0009] For details, please refer to Figure 2, which is a schematic diagram of signature verification process 20. Signature verification process 20 includes the following steps:

[0010] Step 200: Begin.

[0011] Step 201: Use the signature capture device 101 to capture a handwriting data of a signer during the process of making a signature.

[0012] Step 202: Use the image capturing device 100 to capture an image of the signer during the signing process.

[0013] Step 204: Obtain at least one feature information from the image data.

[0014] Step 206: Compare the handwriting data and the at least one feature information with at least one authentication reference data to generate a comparison result.

[0015] Step 208: Based on the comparison result, determine the authenticity of the signature process of the signer.

[0016] Step 210: End.

[0017] According to the signature verification process 20, when the signer signs, the signature verification system 10 captures the signer's handwriting data during the signing process using the signature capture device 101 (step 201). Simultaneously, the signature verification system 10 captures image data of the signer's signing process in real time using the image capture device 100 (step 202), and the processing unit 102 analyzes the image data of the signing process to obtain feature information (step 204). Then, the signature verification system 10 compares the captured handwriting data and feature information with pre-stored authentication reference data (step 206) to determine the authenticity of the signer's signing process (step 208). In short, this embodiment of the invention verifies the authenticity of a signature by performing image analysis and comparison on the handwriting data and image data of the signing process. Since the signing process of a signer contains a variety of static or dynamic information, it can reflect the signer's writing habits, multi-dimensional hand features and other personalized characteristics, making it extremely difficult for others to completely imitate. Compared with traditional signature verification mechanisms that can only verify the signer's handwriting, the embodiments of the present invention can greatly improve the accuracy and reliability of signature verification.

[0018] Specifically, in step 201, the signature verification system 10 captures the signer's handwriting data through the signature capture device 101. The handwriting data may include, but is not limited to, the two-dimensional spatial coordinates of the handwriting, changes in pressure during writing, the tilt angle between the pen tip and the writing plane, and the trajectory of various changes over time, and is not limited to these. For example, in one embodiment, the signature capture device 101 may be implemented using a capacitive touchpad or an electromagnetic induction digitizer, which can detect information such as the pressure, position, and tilt angle of the signer's writing tool (stylus or finger), and can convert the detected information into digital signals. In another embodiment, the signature capture device 101 may be configured with a high-resolution sensing array, which can accurately record the trajectory of the pen tip moving on the plane, including the starting point, ending point, and intermediate path points of the strokes, to fully detect the movement characteristics of the signing process. In another embodiment, the signature capturing device 101 may have pressure sensing capabilities to record changes in pressure applied by the pen tip to the writing surface during the signing process, reflecting variations in the force exerted by the signer across different strokes. In another embodiment, the signature capturing device 101 may record the temporal information of each stroke, including writing speed, acceleration changes, and pause times between strokes, reflecting the signer's writing habits and rhythm. In yet another embodiment, the signature capturing device 101 may also capture the tilt angle between the pen tip and the writing surface, reflecting the signer's pen grip and writing habits, providing additional personal characteristics. This information can be recorded in real-time as the signing process proceeds and converted into digital format for storage as handwriting data, for subsequent feature comparison and verification.

[0019] In step 202, the signature verification system 10 captures image data of the signer's signing process in real time through the image capturing device 100. In this case, the image capturing device 100 should be properly set up to ensure that it can completely capture important information such as the signer's hand movements (relative positions of fingers and knuckles) and signature trajectory (relative movements of fingers and knuckles). The image data can be a continuous dynamic image sequence or a static image captured at a specific time point, and is not limited to these. Furthermore, although the signature verification system 10 in the embodiment of Figure 1 includes only a single image capturing device 100, it is not limited to this, and the signature verification system 10 may also include multiple image capturing devices. For example, in one embodiment, the signature verification system 10 may be equipped with multiple image capturing devices, and the image capturing device with the best shooting angle or range may be selected to generate image data of the signing process; in another embodiment, the signature verification system 10 may integrate the images of multiple image capturing devices to generate image data of the signing process; in yet another embodiment, two or more image capturing devices may be integrated into one device and have a fixed relative position; such techniques for generating the required image data using multiple image capturing devices are skills familiar to those skilled in the art, so detailed operation methods will not be described here.

[0020] In step 204, the processing unit 102 analyzes the image data of the signing process to obtain feature information. In one embodiment, the feature information may be the dynamic trajectory of the signer when signing, such as the distance, path, speed, and acceleration of handprints or knuckle movements, or the distance, path, speed, and acceleration of movements between multiple knuckles. For example, please refer to Figure 3A, which is a schematic diagram of a signing process image captured by the image capturing device 100. According to the signing process image shown in Figure 3A, when the signer is signing, one of his knuckles A moves along a trajectory RT. Therefore, the trajectory RT can reflect the personalized characteristics of the signer when signing, and thus can be used as feature information. The image capturing device 100 can also simultaneously detect the movement trajectories of multiple knuckles and superimpose the trajectories as feature information. In other words, the processing unit 102 can analyze the signing process image captured by the image capturing device 100 in Figure 3A, identify the trajectory RT, and use it as feature information for subsequent comparison.

[0021] In another embodiment, the feature information may be the relative positions of at least three finger joints in three-dimensional space during the signing process, which can reflect the characteristics of the signer's signing posture, such as the relative positions between the metacarpophalangeal joint of the index finger, the proximal phalanx of the index finger, the metacarpophalangeal joint of the middle finger, the proximal phalanx of the middle finger, or the proximal phalanx of the thumb. For example, please refer to Figure 3B, which is a schematic diagram of another signing process image captured by the image capturing device 100. According to the signing process image shown in Figure 3B, when the signer is signing, a specific polygon is formed between multiple finger joints, and this polygon can be defined by distances L1 to L5, reflecting personalized characteristics such as the signer's pen grip or posture when signing, and thus can be used as feature information. In other words, the processing unit 102 can analyze the signing process image captured by the image capturing device 100 in Figure 3B, identify the distances L1 to L5 therein, and use them as feature information for subsequent comparison.

[0022] In another embodiment, the feature information may be the texture features of at least one finger of the signer during the signing process, such as the texture features of multiple knuckles, which can also serve as additional identity verification evidence. For example, please refer to Figure 3C, which is a schematic diagram of another signature process image captured by the image capturing device 100. According to the signature process image shown in Figure 3C, when the signer is signing, multiple knuckles exhibit specific texture patterns C1 to C4. Therefore, texture patterns C1 to C4 and the relationship between them can reflect personalized characteristics such as the signer's pen grip or posture when signing, and thus can be used as feature information. In other words, the processing unit 102 can analyze the signature process image captured by the image capturing device 100 in Figure 3C, identify the texture patterns C1 to C4 therein, and use them as feature information for subsequent comparison.

[0023] In another embodiment, the feature information may be a static posture at a specific point in time during the signing process, such as the hand state at the start of signing, during signing, and when signing is completed, or the position of the lines on the fingers at a specific point in time, or the relative position between at least three finger joints. For example, please refer to Figure 3D, which is a schematic diagram of another signing process image captured by the image capturing device 100. The signing process image shown in Figure 3D is the static posture of the signer at a specific point in time during the signing process, which can reflect the signer's pen grip or posture and other personalized characteristics, and therefore can be used as feature information. In other words, the processing unit 102 can analyze the signing process image captured by the image capturing device 100 in Figure 3D, identify the static posture of the signer during the signing process, and use it as feature information for subsequent comparison.

[0024] Furthermore, it should be noted that the aforementioned signing process is not limited to whether the signer holds a pen to sign. In some applications, the signer may sign with their finger. In such cases, the signature information should be adjusted appropriately according to different applications. This should be a skill familiar to those with ordinary knowledge in the art. In other words, regardless of the signing method used by the signer, as long as signature information reflecting the signer's writing habits can be extracted, it can serve as an important basis for judging the authenticity of the signature, thereby greatly improving security.

[0025] In step 206, the processing unit 102 compares the captured handwriting data and feature information with pre-stored authentication reference data, which may be obtained through various channels. For example, in one embodiment, the signer needs to complete a registration process before signing. For instance, the signer can first capture registration image data of the signer's registration signing process using the image capturing device 100, and obtain registration feature information from it, storing it as authentication reference data. The process of capturing registration image data is not limited to being performed through the image capturing device 100, but can also be completed through another image capturing device, such as a mobile phone, laptop computer, or other electronic device with image capturing capabilities. In other words, the signer should first use the image capturing device 100 or other electronic device with image capturing capabilities to complete the registration process, so that the signature verification system 10 can obtain authentication reference data from the registration image data for comparison in the subsequent signature process.

[0026] It should be noted that the purpose of the registration process is to generate authentication reference data. However, whether to initiate the registration process or the timing and method of its execution can be adjusted appropriately according to different needs. For example, in one embodiment, when a signer is performing a signing process, if the signature verification system 10 finds that the signer's authentication reference data has not yet been established, it can automatically initiate the registration process to immediately establish the authentication reference data, allowing the user to complete the one-time registration process and then continue with the original signing operation. However, in higher security application scenarios, the signature verification system 10 may require the registration process to be reviewed and authorized by an administrator before it can be executed. This authorization mechanism may include measures such as administrator identity verification, record archiving of the approval process, and full monitoring of the registration process. In addition, the signature verification system 10 may also require the signer to provide additional identity verification documents during registration, or to complete multiple registration and signing processes within a specific time period, depending on the usage scenario, in order to establish more complete and reliable authentication reference data.

[0027] In addition to capturing registration image data through image capturing device 100 or another image capturing device to complete the registration process, in another embodiment, authentication reference data may be pre-stored in a database (local or cloud database). The signature verification system 10 can obtain the authentication reference data stored in the database through data transmission when comparison is required (i.e., step 206). Furthermore, to ensure data security, the data that the signature verification system 10 needs to access or use for signature verification (such as image data, feature information, authentication reference data, registration image data, registration feature information, etc.) can be encrypted during transmission and storage, for example, using blockchain technology for encryption. Correspondingly, if the data has undergone encryption, it needs to be decrypted when accessed or used; for example, if the authentication reference data has undergone encryption, the comparison of feature information and authentication reference data in step 206 should include the operation of decrypting the authentication reference data.

[0028] On the other hand, since the feature information reflects the hand features of the signer when signing, it can be the dynamic trajectory of the signature, the angle of the fingers, the texture features, the static posture, etc. Therefore, in step 206, appropriate features can be selected as the basis for comparison. Further, in one embodiment, the signature verification system 10 can assign different weights to each feature according to parameters such as image quality and shooting angle. In another embodiment, the signature verification system 10 can incorporate artificial intelligence technology to learn the correlation between multiple signature processes performed by the signer in different postures, so as to adjust the authentication reference data and thereby improve the accuracy and adaptability of the comparison. For example, artificial intelligence technology can learn the correlation between multiple signature processes performed by the signer in standing, sitting, bending over, using a pen, using the index finger, etc., summarize the characteristics of the signer's signature, and generate or adjust the authentication reference data accordingly to improve the comparison accuracy when the signer signs in different postures.

[0029] As can be seen from the above different embodiments, the feature information obtained from the image data can be dynamic feature information during the signing process or static feature information at a specific point in time during the signing process. In addition, in one embodiment, the signature verification system 10 can also integrate handwriting data and feature information based on a time axis to form a composite information, and then compare the composite information with the authentication reference data to generate a comparison result. That is to say, the signature verification system 10 can synchronize and integrate various types of data in the signing process in a time sequence. In this process, the signature verification system 10 will pair and combine the handwriting data from the signature capturing device 101 and the feature information obtained by the image capturing device 100 according to the time point of their occurrence to form a set of time-related composite information. For example, the handwriting coordinates, pressure value, writing speed and other handwriting data at a certain point in time during the signing process can be paired with the finger angle, knuckle relative position and other feature information captured at the same time. This pairing process can continue throughout the entire signing process, forming a complete time sequence of features from the beginning to the end of the signature. In this way, the signature verification system 10 can capture the dynamic behavioral characteristics of the signer during the signing process, such as analyzing the changes in the signer's finger posture when writing specific strokes, or the coordinated changes in pressure value and hand position when writing at turning points. These temporally correlated features are more personally identifiable and can increase the accuracy of the judgment.

[0030] After forming the composite information, the signature verification system 10 compares the composite information with the feature information. Since the composite information contains a temporal feature sequence, the comparison process must consider not only the numerical similarity of each feature, but also the consistency of their temporal changes. Such derivative changes should be techniques well known to those skilled in the art.

[0031] On the other hand, during the signing process, the signer's hand movements inevitably obscure the handwriting relative to the image capturing device 100, and may vary depending on the pen grip, writing angle, and personal writing habits. In this case, the signature verification system 10 can use the signature capturing device 101 as the main source of handwriting data to ensure that the stroke trajectory during the signing process is completely recorded. At the same time, although the image capturing device 100 may not be able to completely capture all the handwriting due to hand obstruction, this partial obstruction phenomenon can also be regarded as a unique characteristic of the signer, because different pen grip habits and writing postures will present different obstruction patterns in the image. These patterns, together with the visible part of the handwriting, constitute personalized signature characteristics. In particular, when the image capturing device 100 and the signature capturing device 101 maintain a fixed relative position, this characteristic exhibits a high degree of consistency and repeatability.

[0032] Furthermore, the partial handwriting captured by the image capturing device 100 can still be cross-compared with the complete handwriting recorded by the signature capturing device 101. That is, the feature information can include the image-captured handwriting that matches the handwriting data in the image data. This multi-verification mechanism not only improves the accuracy of signature verification but also provides additional security for anti-counterfeiting identification. The micro-behavioral characteristics of the signer during the signing process, such as changes in the pen grip angle, the relative position of the finger joints, and the dynamic changes in handwriting occlusion, can all be accurately recorded and analyzed through this dual-capture method.

[0033] In step 208, the signature verification system 10 determines the authenticity of the signature process based on the comparison results between the handwriting data and / or feature information and the authentication reference data. Furthermore, when the comparison results indicate that the signature is not authentic, the signature verification system 10 can take several measures, such as generating a warning signal, such as a visual cue signal, an audio cue signal, or a tactile feedback signal, or sending a notification message to a designated receiver or administrator to immediately report the abnormal situation or trigger other preset security mechanisms. In addition, the signature verification system 10 can dynamically adjust the judgment threshold value according to different application scenarios to balance the needs of security and convenience.

[0034] In short, through the signature verification process 20, this embodiment of the invention can not only capture the static features of the signature result (such as handwriting), but also capture the dynamic information of the entire signing process, such as the speed and force of the signature, habitual pauses when writing, and the stroke order when finishing the stroke. These habitual actions are often formed over a long period of time and are unconscious, making them extremely difficult for others to completely imitate. In addition to writing habits, this embodiment of the invention can also capture multi-dimensional features such as the signer's hand movements, signature trajectory, and finger angles. The combination of these dynamic features further increases the difficulty of forging signatures and greatly improves the accuracy and reliability of verification. Furthermore, this embodiment of the invention can integrate handwriting data and feature information based on a timeline to form composite information, thus enabling a more comprehensive understanding of the signer's signature characteristics, including not only static handwriting shapes but also temporally changing behavioral features, thereby improving the reliability and security of signature verification. This embodiment of the invention can also encrypt and decrypt the data that needs to be accessed or used for signature verification to ensure data security. Furthermore, embodiments of the present invention can also integrate artificial intelligence technology. Through machine learning algorithms, the signature verification system 10 can learn and recognize the signature characteristics of the same person in different situations, such as signature methods using different writing tools or in different postures, thereby improving the accuracy and adaptability of verification. Simultaneously, the signature verification system 10 of the present invention can be implemented on various devices with image capture capabilities or electronic devices capable of connecting to external image capture devices, including but not limited to personal computers, laptops, or mobile devices.

[0035] Therefore, the signature verification process 20 of this embodiment of the invention can ensure the reliability and security of signature verification through multiple feature analysis, flexible data sources, encryption protection mechanisms, artificial intelligence assistance, and a comprehensive warning system, and can effectively prevent signature forgery, and also provides sufficient flexibility and scalability for practical applications.

[0036] Furthermore, it should be noted that the signature verification process 20 represents the main operating mode of the signature verification system 10. When implementing the signature verification system 10, those skilled in the art should select appropriate components to correctly execute each step of the signature verification process 20 or its derivative variations. Specifically, the processing unit 102 can be a microprocessor, a digital signal processor (DSP), or a microcontroller, and the storage unit 104 can be a read-only memory (ROM), random access memory (RAM), flash memory, or other types of memory devices. The program code 106 is stored in the storage unit 104. When the system starts, the processing unit 102 can read and execute the program code 106 to realize the various functions of the signature verification process 20. In another embodiment, some computationally intensive or time-sensitive functions can be implemented using hardware circuits, while other more complex logical judgments can be handled through software programs. This hybrid implementation strikes a balance between performance and flexibility, ensuring both system immediacy and good scalability. Regardless of the implementation method, the processing unit 102 must correctly acquire the signals or data captured by the image capturing device 100, which can be achieved, for example, through standard communication protocols such as Serial Communication Interface (SCI), Serial Peripheral Interface (SPI), or Inter-Integrated Circuit (I2C).

[0037] On the other hand, the signature verification system 10 uses the image capturing device 100 to capture image data of the signer's signing process, and performs image analysis and comparison to verify the authenticity of the signature. Therefore, the selection or setting of the image capturing device 100, such as its placement, angle, sensitivity, focal length, resolution, and other parameters, should prioritize the ability to clearly capture the signer's hand movements and the signing process. In addition, to ensure that the signer signs in the appropriate position, a specific signature position can be planned in advance based on the shooting characteristics of the image capturing device 100, or it can be integrated with the signature capturing device 101 so that the signer can sign in the appropriate position. Furthermore, the signature verification system 10 can also incorporate a prompting mechanism, such as issuing sound or flashing lights when the signer does not sign in the appropriate position to prompt the signer to adjust the signature position.

[0038] The signature capture device 101 is responsible for capturing the handwriting trajectory data of the signer when signing. It can be achieved through high-precision pressure sensing technology, touch screen or other sensing elements to capture the stroke changes, writing speed and pressure differences during the signing process in real time and accurately, and convert them into digital handwriting data for subsequent verification and analysis. Therefore, the selection of the signature capture device 101 should take into account its sensitivity, resolution and response speed to writing actions to ensure that it can accurately capture all the writing details of the signer. In addition, the signature capture device 101 can be designed with an ergonomic shape for easy operation by the signer, and can be equipped with a haptic feedback system or display interface to help the signer confirm whether their signature action meets the requirements.

[0039] The signature verification system 10 or signature verification process 20 is designed to verify the authenticity of signatures. Those skilled in the art should be able to appropriately apply or implement it in various scenarios requiring identity verification. For example, in a corporate office environment, the signature verification system 10 or signature verification process 20 can be applied to an electronic document signing system to ensure the authenticity of the signatory's identity and can be integrated into the enterprise's information security mechanism as an identity verification method for employees to log in to workstations or access confidential information. In the financial sector, the signature verification system 10 or signature verification process 20 can be applied to identity verification in bank counter services to improve transaction security; it can also be integrated into mobile banking applications to replace traditional password verification, providing users with a more secure and convenient login method. In the retail sector, the signature verification system 10 or signature verification process 20 can be applied to identity verification in electronic payment systems, especially for authorization confirmation of large transactions; it can also be used in membership card systems, allowing members to quickly complete identity verification through personalized signature methods. In the healthcare field, the signature verification system 10 or signature verification process 20 can be used for electronic signing of medical records and prescriptions, ensuring the authenticity and integrity of medical documents; it can also be applied to medical personnel's attendance check-in or operating room access management. In the education field, the signature verification system 10 or signature verification process 20 can be used for identity verification on distance learning platforms, ensuring the identity of students participating in online exams; it can also be used for library borrowing systems or laboratory equipment usage management. In property management, the signature verification system 10 or signature verification process 20 can be used for resident identity verification when entering and exiting buildings, providing higher security than traditional access cards; it can also be used for registration of various public facilities. In the logistics field, the signature verification system 10 or signature verification process 20 can be used for package receipt confirmation, not only recording the recipient's signature but also instantly verifying the recipient's identity. In public sector services, the signature verification system 10 or signature verification process 20 can be used for electronic signing of various government documents, improving administrative efficiency while ensuring the legal validity of documents. These diverse application scenarios demonstrate that this invention can provide a safer and more convenient solution for identity verification needs across various industries.

[0040] In summary, the signature verification system or process provided by the present invention captures dynamic features during the signing process in real time through an image capturing device. It not only includes the static information required for traditional signature comparison but also records the signer's unique writing habits and movement characteristics. During the analysis and comparison of feature information, the present invention integrates multiple verification mechanisms, including dynamic trajectory analysis, finger angle recognition, and biometric comparison, and can continuously learn and optimize judgment criteria through artificial intelligence technology. Furthermore, the present invention provides comprehensive protection measures for data security, including encrypted data storage and anomaly alert mechanisms, ensuring the reliability of the entire verification process. Therefore, the signature verification system or process of the present invention can not only effectively prevent signature forgery but also meet the needs of various practical application scenarios, providing a practical and secure solution for identity verification in the digital age. The above are merely preferred embodiments of the present invention. All equivalent changes and modifications made within the scope of the claims of the present invention should be considered within the scope of the present invention. [Simplified Explanation of the Diagram]

[0041] Figure 1 is a schematic diagram of the signature verification system according to Embodiment 1 of the present invention. Figure 2 is a schematic diagram of the signature verification process according to Embodiment 1 of the present invention. Figures 3A, 3B, 3C and 3D are schematic diagrams of the signature process images.

Claims

1. A signature verification method based on image analysis, comprising the following steps: capturing handwriting data of a signer during a signature process using a signature capturing device; capturing image data of the signer during the signature process using an image capturing device; obtaining at least one feature information from the image data; integrating the handwriting data and the at least one feature information based on a timeline to form a composite information, comparing the composite information with at least one pre-established authentication reference data to generate a comparison result; and determining the authenticity of the signer's signature process based on the comparison result.

2. The signature verification method as described in claim 1, wherein the at least one feature information includes at least one of the following: the angle of at least one finger of the signer when performing the signing process; the texture features on the at least one finger of the signer when performing the signing process; and the relative positions of at least three finger joints of the signer when performing the signing process.

3. The signature verification method as described in claim 2, wherein the at least one feature information is a dynamic feature information in the signature process or a static feature information at a specific point in time in the signature process.

4. The signature verification method as described in claim 1, wherein the at least one feature information includes: a partially occluded handwriting obtained by comparing the image data with the handwriting data.

5. A signature verification system based on image analysis, comprising: a signature capturing device for capturing handwriting; an image capturing device for capturing images; a processing unit coupled to the signature capturing device and the image capturing device; and a storage unit coupled to the processing unit and storing code that instructs the processing unit to execute a signature verification method, the signature verification method comprising the following steps: capturing handwriting data of a signer during a signature process using the signature capturing device; capturing image data of the signer during the signature process using the image capturing device; obtaining at least one feature information from the image data; integrating the handwriting data and the at least one feature information based on a timeline to form a composite information, comparing the composite information with at least one pre-established authentication reference data to generate a comparison result; and determining the authenticity of the signer's signature process based on the comparison result.

6. The signature verification system as described in claim 5, wherein the at least one feature information includes at least one of the following: the angle of at least one finger of the signer when performing the signing process; the texture features on the at least one finger of the signer when performing the signing process; and the relative positions of at least three knuckles of the signer when performing the signing process.

7. The signature verification system as described in claim 5, wherein the at least one feature information is a dynamic feature information in the signing process or a static feature information at a specific point in time in the signing process.

8. The signature verification system as described in claim 5, wherein the at least one feature information includes: a partially obscured handwriting obtained by comparing the image data with the handwriting data.

Images