技术领域technical field
本发明涉及一种基于图像特征融合与图像识别的签名鉴伪系统及方法，属于数字图像处理与笔迹鉴定技术领域。The invention relates to a signature verification system and method based on image feature fusion and image recognition, and belongs to the technical field of digital image processing and handwriting verification.
背景技术Background technique
信息技术的飞速发展给人们的日常生活带来了极大的便利，与此同时对个人身份进行准确的认证、保护信息安全则成了当今信息化时代亟需解决的一个关键问题。手写签名是人的一种比较稳定的行为特征，利用手写签名进行个人身份的认证具有非侵犯性(或非触性)、易于获取、容易使人接受等特点，是一种重要的个人身份的表示手段。The rapid development of information technology has brought great convenience to people's daily life. At the same time, accurate authentication of personal identity and protection of information security have become a key problem that needs to be solved urgently in today's information age. Handwritten signature is a relatively stable behavioral characteristic of people. Using handwritten signature for personal identity authentication is non-invasive (or non-contact), easy to obtain, and easy to accept. It is an important personal identity. Express means.
手写签名鉴伪隶属于笔迹鉴定这一技术领域，传统的基于人工的笔迹鉴定在实际操作中暴露出了种种弊端和缺陷：如鉴定机构缺乏相互配合机制，鉴定人员资格认定标准不规范、鉴定水平参差不齐难以保障鉴定的正确性等等都会对鉴定结果造成负面影响。因此利用计算机自动化、规范化的方式来处理笔迹鉴定这一原本非工程的领域，对该领域的发展具有很大的推动作用。Handwritten signature authentication belongs to the technical field of handwriting identification. The traditional manual handwriting identification has exposed various drawbacks and defects in actual operation: for example, the identification institutions lack mutual cooperation mechanism, the qualification standards of identification personnel are not standardized, and the level of identification Unevenness makes it difficult to guarantee the correctness of the identification, etc., which will have a negative impact on the identification results. Therefore, the use of computer automation and standardized methods to deal with handwriting identification, which was originally a non-engineering field, has a great role in promoting the development of this field.
目前签名鉴伪按实际操作方式的不同可分为联机和脱机两种，两者都有很广的应用背景，可在诸如金融、保险、公安司法部门的刑事调查和法庭审判等领域发挥重要作用。这些技术具有鉴定快、效率高、不受文检人员主观因素的影响等特点。从实际来讲联机手写签名的鉴伪技术已经十分成熟并已进入实用化阶段，然而脱机签名因无法像联机签名获取到签名者书写时的笔画顺序、书写速度、运笔压力等动态信息，无疑在鉴定真伪的难度上会更大，所以脱机签名的鉴定技术目前还不太成熟，但也成为了当下的研究热点。At present, signature verification can be divided into two types: online and offline according to different actual operation methods. Both of them have a wide application background and can play an important role in fields such as finance, insurance, criminal investigation and court trial of public security and judicial departments. effect. These technologies have the characteristics of quick identification, high efficiency, and not being affected by the subjective factors of document inspectors. From a practical point of view, the counterfeiting technology of online handwritten signatures is very mature and has entered the practical stage. However, offline signatures cannot obtain dynamic information such as the order of strokes, writing speed, and pen pressure of the signer when writing like online signatures. It will be more difficult to verify the authenticity, so the verification technology of offline signature is not yet mature, but it has become a current research hotspot.
发明内容Contents of the invention
本发明针对当前社会热门的手写签名鉴定中的技术空缺，提出了一种基于图像特征融合与图像识别的签名鉴伪系统及方法，试图改善脱机签名鉴伪的现状。Aiming at the technical vacancy in the currently popular handwritten signature identification in the society, the present invention proposes a signature authentication system and method based on image feature fusion and image recognition, and tries to improve the current situation of offline signature authentication.
一种基于图像特征融合识别的签名鉴伪系统，其特征是，主要包括签名图像数据库模块、图像预处理模块、特征向量提取模块、特征向量融合模块以及特征相似性度量模块；A signature authentication system based on image feature fusion recognition, characterized in that it mainly includes a signature image database module, an image preprocessing module, a feature vector extraction module, a feature vector fusion module and a feature similarity measurement module;
还包括图像录入模块和结果输出反馈模块；It also includes an image input module and a result output feedback module;
签名图像数据库模块预先存储注册用户的个人手写签名图像、个人信息，后续的鉴伪过程中调用这些信息来核实待鉴定的签名图像的真伪；The signature image database module pre-stores the registered user's personal handwritten signature image and personal information, and calls these information in the subsequent authentication process to verify the authenticity of the signature image to be authenticated;
图像预处理模块对录入进系统的待鉴定签名图像进行预处理；The image preprocessing module preprocesses the image of the signature to be authenticated entered into the system;
特征向量提取模块对经预处理模块处理之后的待鉴定签名图像进行特征提取；The feature vector extraction module performs feature extraction on the signature image to be authenticated after being processed by the preprocessing module;
特征向量融合模块采用特征层融合模型对特征向量提取模块提取的特征进行融合；The feature vector fusion module adopts the feature layer fusion model to fuse the features extracted by the feature vector extraction module;
特征相似性度量模块：采用夹角余弦法计算待鉴伪签名图像与数据库中预存的签名图像的相似度，通过计算待鉴定签名图像所提取的特征向量与数据库中对应图像的特征向量之间的夹角余弦，判断两者之间的相似度，作为鉴定真伪的评判依据。Feature similarity measurement module: use the angle cosine method to calculate the similarity between the signature image to be authenticated and the signature image pre-stored in the database, and calculate the difference between the feature vector extracted from the signature image to be authenticated and the feature vector of the corresponding image in the database The cosine of the included angle is used to judge the similarity between the two as the basis for judging the authenticity.
所述图像预处理模块的预处理过程包括灰度化处理、二值化处理以及大小归一化处理的步骤。The preprocessing process of the image preprocessing module includes the steps of gray scale processing, binarization processing and size normalization processing.
特征向量提取模块中包括比例特征提取子模块、纹理特征提取子模块、弹性网格特征提取子模块，分别分析并提取签名图像的比例特征、纹理特征和弹性网格特征；The feature vector extraction module includes a proportional feature extraction sub-module, a texture feature extraction sub-module, and an elastic grid feature extraction sub-module, respectively analyzing and extracting the proportional feature, texture feature and elastic grid feature of the signature image;
比例特征提取子模块：提取签字人每个签名字符的细化后的笔画长度与所占面积的比例作为特征值数据；Proportional feature extraction sub-module: extract the ratio of the thinned stroke length and occupied area of each signature character of the signatory as feature value data;
纹理特征提取子模块：提取反映字符的字型、字位倾斜、笔划方向、笔划和部首搭配的纹理特征；Texture feature extraction sub-module: extract the texture features that reflect the character font, font position inclination, stroke direction, stroke and radical collocation;
弹性网格特征提取子模块：利用一组设想的网线对签名图像进行区域划分，使任意两条相邻网线之间的目标像素个数相等；将字体按横、竖、撇、捺四方向分解后，提取横、竖、撇、捺四个方向特征向量值结合在一起，构成一个完整的字体特征矢量。Elastic grid feature extraction sub-module: use a group of imaginary network lines to divide the signature image into regions, so that the number of target pixels between any two adjacent network lines is equal; decompose the font in four directions: horizontal, vertical, left and right Finally, extract the four direction feature vector values of horizontal, vertical, left and right and combine them together to form a complete font feature vector.
一种基于图像特征融合识别的签名鉴伪方法，其特征是：包括以下步骤：A method for authenticating signatures based on fusion and identification of image features is characterized in that it comprises the following steps:
扫描仪完成对需鉴伪签名图像的信息采集；The scanner completes the information collection of the signature image to be authenticated;
特征向量提取模块提取需鉴伪签名图像的比例特征、纹理特征和弹性网格三大特征；The feature vector extraction module extracts the three major features of the signature image that needs to be authenticated: scale feature, texture feature and elastic grid;
特征向量融合模块将提取的三大特征融合；The feature vector fusion module fuses the extracted three major features;
采用夹角余弦法计算待鉴伪签名图像与样本签名图像的相似度，通过计算待鉴定签名图像所提取的特征向量与样本对应图像的特征向量之间的夹角余弦，以余弦值衡量相似程度，输出鉴伪结果。Using the angle cosine method to calculate the similarity between the signature image to be authenticated and the sample signature image, by calculating the angle cosine between the feature vector extracted from the signature image to be authenticated and the feature vector of the sample corresponding image, the similarity is measured by the cosine value , output the result of forgery verification.
包括如下详细步骤：将需鉴伪签名图像运用扫描仪录入至特征融合识别的签名鉴伪系统，根据人物信息从数据库中调出其原始签名图像的特征数据；It includes the following detailed steps: use a scanner to input the signature image to be authenticated into the signature authentication system for feature fusion recognition, and call out the feature data of the original signature image from the database according to the person information;
对扫描的图片预处理得到归一化的二值图像，进行二值图像的特征检测和分析；Preprocess the scanned image to obtain a normalized binary image, and perform feature detection and analysis of the binary image;
提取比例、纹理和弹性网格三大特征，同时进行特征值向量化，再将三大特征进行融合后和原始签名图像的特征数据通过向量夹角余弦法相比较，得出鉴伪结果。The three major features of scale, texture and elastic grid are extracted, and the eigenvalues are vectorized at the same time, and then the three major features are fused and compared with the feature data of the original signature image through the vector angle cosine method to obtain the counterfeit verification result.
比例特征的提取方法是通过计算签名二值图像中比例特性作为特征值，分别对签名的字符逐个进行边缘检测，得到相应的矩形框(字符数量和矩形框数量相等，矩形框数目至少大于1，否则重新录入)，计算每个字符的面积大小S，再将需鉴别签名图像进行字体细化处理，将每个字符的像素点进行计数，作为字长L；将每个字符L/S的数据作为二位向量组的数据，建立一个向量组，得到比例特性。The method for extracting the proportional feature is to calculate the proportional feature in the signature binary image as the feature value, respectively carry out edge detection on the characters of the signature one by one, and obtain the corresponding rectangular frame (the number of characters is equal to the number of rectangular frames, and the number of rectangular frames is at least greater than 1, Otherwise, re-enter), calculate the area size S of each character, and then perform font refinement processing on the signature image to be identified, and count the pixels of each character as the word length L; the data of each character L/S As the data of the two-bit vector group, a vector group is established to obtain the proportional characteristic.
纹理特征的提取方法，通过对签名图像纹理特性的提取，运用Gabor滤波器提取纹理特性；Gabor函数在空间域和频率域中同时进行测量，主要目标为签名笔画具有一定的线条宽度和方向，每幅笔迹样本图像经每一通道滤波即提取笔迹纹理的特征，在样本图像I(x,y)中抽取样点(X,Y)，该点处提取的特征通过计算各个通道滤波器后图像的均值和标准差作为特征组成特征向量，按照特征值位置，经行特征值向量组化，得到n维特征值数据向量组。The method of extracting the texture features, through the extraction of the texture features of the signature image, uses the Gabor filter to extract the texture features; the Gabor function is measured in the space domain and the frequency domain at the same time, and the main goal is that the signature strokes have a certain line width and direction. A handwriting sample image is filtered by each channel to extract the features of the handwriting texture, and the sample point (X, Y) is extracted in the sample image I (x, y), and the features extracted at this point are calculated by calculating the image after each channel filter. The mean and standard deviation are used as features to form a feature vector, and according to the position of the feature value, the feature value vector is grouped to obtain an n-dimensional feature value data vector group.
弹性网格的提取方法，通过对签名图像弹性网格特征值提取比对，采用弹性网格，利用二值图像，当所取标志量为1时表示为黑像素点，当所取标志量为0时表示为白像素点；然后将字体横、竖、撇、捺四方向分解后，提取横方向汉字子图像fH(x,y)，横方向第i个网格内特性运用特征值计算公式计算得到，其他几个方向的特征同理，横、竖、撇、捺四个方向特征值结合在一起，构成一个完整的字体特征向量。The extraction method of the elastic grid, by extracting and comparing the eigenvalues of the elastic grid of the signature image, using the elastic grid, using the binary image, when the taken sign is 1, it is represented as a black pixel, when the taken sign is 0 Represented as white pixels; then decompose the font in the four directions of horizontal, vertical, left and right, and extract the sub-image fH (x, y) of Chinese characters in the horizontal direction, and use the eigenvalue calculation formula to calculate the characteristics of the i-th grid in the horizontal direction Obtained, the characteristics of the other directions are the same, the eigenvalues of the four directions of horizontal, vertical, left and right are combined to form a complete font feature vector.
签名鉴伪数据对比工作主要依据原始的数据库，数据库的构建方法是提前录入需检测签名的人员信息，主要包括以上三大特征性数据相关特征值向量组的存储。The signature authentication data comparison work is mainly based on the original database. The database construction method is to enter the information of the personnel who need to detect the signature in advance, mainly including the storage of the above three characteristic data-related eigenvalue vector groups.
签名鉴伪系统通过特征提取分析模块得到三大模块的特征向量，运用特征融合技术将三大模块的特征值进行融合后，存储至数据库或是备份进行相似性度量工作。The signature verification system obtains the feature vectors of the three modules through the feature extraction and analysis module, uses the feature fusion technology to fuse the feature values of the three modules, and stores them in the database or backup for similarity measurement.
本发明所达到的有益效果：The beneficial effect that the present invention reaches:
本发明公开了一种基于图像特征融合与图像识别的签名鉴伪系统及方法，可以进行脱机签名鉴伪。通过提取比例特征、纹理特征、弹性网格特征作为三大主要特征参数，采用特征融合模型将提取的特征进行融合，以特征向量夹角余弦作为相似性度量的主要参量，完成签名鉴伪工作，鉴伪结果稳定、客观。The invention discloses a signature authentication system and method based on image feature fusion and image recognition, which can perform offline signature authentication. By extracting scale features, texture features, and elastic grid features as the three main feature parameters, the feature fusion model is used to fuse the extracted features, and the cosine of the angle between feature vectors is used as the main parameter of similarity measurement to complete the signature authentication work. The results of counterfeiting are stable and objective.
附图说明Description of drawings
图1为系统模块结构图。Figure 1 is a block diagram of the system.
图2为特征层融合模型图。Figure 2 is a diagram of the feature layer fusion model.
图3为系统处理过程图。Figure 3 is a process diagram of the system processing.
图4为系统鉴定为真实签名的界面图。Fig. 4 is an interface diagram of the system authenticating the signature.
图5为系统鉴定为伪造签名的界面图。Fig. 5 is an interface diagram of the system identifying a forged signature.
图6为数据库的建立示例(可根据编号、姓名、年龄、身份证号等信息搜寻原始数据)。Figure 6 is an example of database establishment (the original data can be searched according to information such as number, name, age, ID number, etc.).
图7为系统测试结果统计表。Figure 7 is a statistical table of system test results.
具体实施方式Detailed ways
本系统采用了模块化的设计方法，整个系统主要由签名图像数据库模块、图像预处理模块、特征向量提取模块、特征向量融合模块以及相似性度量模块这五大模块组成，辅助模块则包括了图像录入模块和结果输出反馈模块(如图1所示)，具体方案如下：This system adopts a modular design method. The whole system is mainly composed of five modules: signature image database module, image preprocessing module, feature vector extraction module, feature vector fusion module and similarity measurement module. The auxiliary module includes image input Module and result output feedback module (as shown in Figure 1), the specific scheme is as follows:
(1)签名图像数据库模块预先存储了注册用户的个人手写签名图像、个人信息(如姓名、性别、身份证号码等)等数据化信息，如图6，后续的鉴伪过程中可以调用这些信息来核实待鉴定的签名图像的真伪；(1) The signature image database module pre-stores digital information such as the registered user's personal handwritten signature image, personal information (such as name, gender, ID number, etc.), as shown in Figure 6, and these information can be called in the subsequent authentication process To verify the authenticity of the signature image to be identified;
(2)图像预处理模块主要完成对录入进系统的待鉴定签名图像进行预处理的工作，预处理的过程具体包括了灰度化处理、二值化处理以及大小归一化处理这几个具体步骤，将预处理之后的签名图像再用于后续模块的处理；(2) The image preprocessing module mainly completes the preprocessing of the image of the signature to be authenticated entered into the system. The preprocessing process specifically includes grayscale processing, binarization processing, and size normalization processing. step, using the preprocessed signature image for subsequent module processing;
(3)特征向量提取模块是对经预处理模块处理之后的待鉴伪签名图像进行特征提取的模块，又可细分为三个子模块，各个模块分别分析并提取了签名图像的三大特征：比例特征、纹理特征和弹性网格特征。(3) The feature vector extraction module is a module that extracts features from the forged signature image to be authenticated after being processed by the preprocessing module, and can be subdivided into three sub-modules. Each module analyzes and extracts three major features of the signature image: Scale features, texture features, and elastic mesh features.
①比例特征提取子模块：提取签字人每个签名字符的细化后的笔画长度与所占面积的比例作为特征值数据。① Proportional feature extraction sub-module: extract the ratio of the thinned stroke length and occupied area of each signature character of the signatory as feature value data.
②纹理特征提取子模块：纹理特征能够反映字符的字型、字位倾斜、笔划方向、笔划和部首搭配这些常用的、比较稳定的、鉴别能力强的特征，本子模块使用Gabor滤波器提取签名图像的纹理特征。②Texture Feature Extraction Sub-module: Texture feature can reflect the character font, word position inclination, stroke direction, stroke and radical collocation, which are commonly used, relatively stable, and strong identification features. This sub-module uses Gabor filter to extract signatures Image texture features.
③弹性网格特征提取子模块：该模块利用一组假想的网线对签名图像进行区域划分，使任意两条相邻网线之间的目标像素个数相等。通常是由纵横的直线组成网格。该子模块提取的特征能很好地解决手写体汉字中因书写风格不同引起的笔划位置不稳定、字体局部变形等问题，并且能够有效地反映手写签名的结构细节。③Elastic grid feature extraction sub-module: This module uses a set of imaginary network lines to divide the signature image into regions, so that the number of target pixels between any two adjacent network lines is equal. A grid is usually composed of vertical and horizontal straight lines. The features extracted by this sub-module can well solve the problems of unstable stroke positions and local deformation of fonts caused by different writing styles in handwritten Chinese characters, and can effectively reflect the structural details of handwritten signatures.
(4)特征向量融合模块：该模块采用特征层融合模型(如图2所示)对上述提取的三大特征进行融合。特征层融合是从原始信息中提取特征信息进行综合分析和处理的中间层次过程。所提取的特征信息是原数据层融合原始信息的充分表示量或统计量，并据此对多源信息进行分类、汇集和综合，同时多特征提取可以提供比单特征提取更多的待检测目标的特征信息，从而增大特征空间维数。简言之，特征层融合就是特征层的联合识别，可以有效的改善鉴伪的性能。(4) Feature vector fusion module: this module fuses the three extracted features above using a feature layer fusion model (as shown in Figure 2). Feature level fusion is an intermediate level process of extracting feature information from original information for comprehensive analysis and processing. The extracted feature information is a sufficient representation or statistic of the fusion of the original information of the original data layer, and based on this, the multi-source information is classified, collected and synthesized. At the same time, multi-feature extraction can provide more targets to be detected than single-feature extraction. feature information, thereby increasing the dimensionality of the feature space. In short, feature layer fusion is the joint identification of feature layers, which can effectively improve the performance of counterfeiting.
(5)特征相似性度量模块：该模块采用夹角余弦法计算待鉴伪签名图像与数据库中预存的签名图像的相似度。夹角余弦相似性，也称为余弦相似度或者余弦距离，是用向量空间中两个向量夹角的余弦值作为衡量两个个体间差异大小的度量。向量是多维空间中的有向线段，如果两个向量的方向一致，即夹角接近于零，那么这两个向量就相近，而要确定两个向量的方向是否一致，这就要用到余弦定理计算向量的夹角。在二维空间中向量A(x1,y1)和向量B(x2,y2)的夹角余弦公式：(5) Feature similarity measurement module: This module uses the angle cosine method to calculate the similarity between the signature image to be authenticated and the signature image pre-stored in the database. Angle cosine similarity, also known as cosine similarity or cosine distance, uses the cosine value of the angle between two vectors in a vector space as a measure of the difference between two individuals. A vector is a directed line segment in a multidimensional space. If the directions of two vectors are the same, that is, the angle between them is close to zero, then the two vectors are close. To determine whether the directions of the two vectors are the same, you need to use cosine Theorem computes the angle between vectors. The cosine formula of the angle between vector A(x1 ,y1 ) and vector B(x2 ,y2 ) in two-dimensional space:
通过计算待鉴定签名图像所提取的特征向量与数据库中对应图像的特征向量之间的夹角余弦，即可计算出两者之间的相似度，进而可作为鉴定真伪的评判依据。By calculating the cosine of the angle between the feature vector extracted from the signature image to be authenticated and the feature vector of the corresponding image in the database, the similarity between the two can be calculated, which can then be used as a basis for judging authenticity.
系统实际操作时签名实体(连同签名纸张)垂直摆放于扫描仪下方，要求尽量采取光照适中，录入器材像素相对稳定较高，所用的笔、纸质量较高且统一，纸张尽量选择纯白色无花纹，笔选用0.5mm黑色签字笔(签名字符的细化环节可保证系统对不同签字笔的适应性)。拍摄室中至少使用200W的LED灯为拍摄提供光源，扫描过程中不会出现反光和阴影等干扰因素，从而提高了系统判断分析的准确性。扫描完成后，将图像传至系统进行识别分析，判断后得出反馈。主要的核心流程为脱机签名图像特征值的提取融合与比对分析，在特征值提取比对方面，主要提取了比例特征、纹理特征、弹性网格特征作为三大主要特征参数，将三大特征融合得出特征向量组，运用特征向量夹角余弦作为相似性度量的主要参量，完成签名特征值比对工作，鉴别签名字符的真伪性。签名验证的书写人的签字相关信息是经过注册的，而且登记的每一个人都附有一个身份号码，验证时只需根据待验证者申明的身份调出相应的参考签名特征值数据进行鉴别即可。During the actual operation of the system, the signature entity (together with the signature paper) is placed vertically under the scanner. It is required to use moderate lighting as much as possible. The pixel input equipment is relatively stable and high. The quality of the pen and paper used is high and uniform. 0.5mm black signature pen is selected for the pattern and pen (the refinement of signature characters can ensure the adaptability of the system to different signature pens). At least 200W LED lamps are used in the shooting room to provide light sources for shooting, and there will be no interference factors such as reflections and shadows during the scanning process, thereby improving the accuracy of system judgment and analysis. After the scanning is completed, the image is sent to the system for identification and analysis, and feedback is obtained after judgment. The main core process is the extraction, fusion and comparison analysis of the feature values of the offline signature image. In terms of feature value extraction and comparison, the scale feature, texture feature, and elastic grid feature are mainly extracted as the three main feature parameters. The eigenvector group is obtained by feature fusion, and the cosine angle between the eigenvectors is used as the main parameter of similarity measurement to complete the comparison of signature eigenvalues and identify the authenticity of signature characters. The signature-related information of the person who wrote the signature verification is registered, and each registered person is attached with an identity number. When verifying, it is only necessary to call out the corresponding reference signature characteristic value data for identification according to the identity of the person to be verified. Can.
本发明的系统算法的具体流程如图3所示：The concrete flow of system algorithm of the present invention is as shown in Figure 3:
(1)当签名的真伪需判别时，利用扫描仪对目标图像进行采样，在经过模数转换后，存放于该模块内的图像采集数据存储区中。(1) When the authenticity of the signature needs to be discriminated, use the scanner to sample the target image, and store it in the image acquisition data storage area in the module after analog-to-digital conversion.
(2)设存放在图像采集数据存储区中的彩色图像为I1，复制一份存放于数据存储区并对其进行图像灰度化以得到灰度图像I2，并对I2使用中值滤波以去除噪声，然后使用最大类间方差法(otsu算法)对I2进行阈值分割，得到目标签字二值化图像I3，最终对I3进行大小归一化处理以获得其归一化图像I4，将签字彩色图像I1，原始签字灰度图像I2、二值化处理后的目标签字图像I3、归一化图像I4均暂时保存于数据存储区用于后续处理。(2) Let the color image stored in the image acquisition data storage area be I1 , copy a copy and store it in the data storage area and grayscale the image to obtain the grayscale image I2 , and use the median value for I2 Filter to remove noise, then use the maximum between-class variance method (otsu algorithm) to thresholdI2 to obtain the target signature binarized imageI3 , and finally perform size normalization onI3 to obtain its normalized image I4 , temporarily save the signature color image I1 , the original signature grayscale image I2 , the binarized target signature image I3 , and the normalized image I4 in the data storage area for subsequent processing.
(3)将大小归一化的二值图像I4中汉字的大小提取出来作为特征比例特性，首先确定每行中各个文字的左右边框计算投影图hist(i)与横坐标的交点，记录这些交点所在的列坐标分别为d1、d2和d3，用数组依次存储起来，分别对d1和d2、d2和d3之间的黑色像素进行统计，若和大于零，即可认为d1和d2、d2和d3分别为签名第一个字和第二个字、等二个字与第三个字的左右边框。再预设两个标志变量jug1＝0和jug2＝0，对每行中每个字符通过由上而下的顺序查询。第一次搜索到黑色像素置jug1＝1，此时可根据该黑色像素点确定上边框所在的行即纵坐标。然后通过由下而上的顺序查询，第一次搜索黑色像素点并置jug2＝1，得到下边框所在的行即纵坐标。从而确定了每个字的上下边框。确定边框之后，计算每个字符的面积大小S。再将需鉴别签名图片进行字体细化处理，将每个字符的像素点进行计数，作为字长L。将每个字符L/S的数据作为二位向量组的数据，建立一个向量组，从而得到比例特性。(3) Extract the size of the Chinese characters in the binary image I4 of size normalization as the feature ratio feature, first determine the left and right borders of each character in each line, calculate the intersection point of the projection map hist(i) and the abscissa, and record these The column coordinates where the intersection points are d1, d2, and d3 are respectively stored in an array, and the black pixels between d1 and d2, d2 and d3 are counted respectively. If the sum is greater than zero, it can be considered that d1 and d2, d2 and d3 are the left and right borders of the first and second characters, the second character and the third character of the signature respectively. Two flag variables jug1=0 and jug2=0 are preset, and each character in each line is queried in order from top to bottom. Set jug1=1 when a black pixel is found for the first time, and at this time, the row where the upper frame is located can be determined according to the black pixel point, that is, the vertical coordinate. Then, through sequential query from bottom to top, search for black pixels for the first time and set jug2=1 to get the row where the lower border is located, that is, the vertical coordinate. Thus, the upper and lower borders of each word are determined. After determining the border, calculate the area size S of each character. Then, the image of the signature to be authenticated is subjected to font refinement processing, and the pixels of each character are counted as the word length L. The data of each character L/S is used as the data of the two-bit vector group, and a vector group is established to obtain the proportional characteristic.
(4)纹理特性的提取，运用Gabor滤波器提取纹理特性，Gabor变换在分析数字图像中局部区域的频率和方向信息方面具有优异的性能，即它能做到时域信号和频域信号的局部化。Gabor函数可在空间域和频率域中同时进行测量，并且在这两种域中都是局部的变换，具有明显的方向选择性和频率特性。由于签名笔画具有一定的线条宽度和方向，首先从签名图像的统计信息入手，每幅笔迹样本图像经每一通道滤波即提取笔迹纹理的特征，在样本图像I(x,y)中抽取样点(X,Y)，则在该点处提取的特征Z为：(4) Extraction of texture characteristics, using Gabor filter to extract texture characteristics, Gabor transform has excellent performance in analyzing the frequency and direction information of local areas in digital images, that is, it can achieve local time domain signals and frequency domain signals change. The Gabor function can be measured simultaneously in the space domain and the frequency domain, and it is a local transformation in both domains, with obvious direction selectivity and frequency characteristics. Since the signature strokes have a certain line width and direction, firstly start with the statistical information of the signature image, each handwriting sample image is filtered by each channel to extract the characteristics of the handwriting texture, and sample points are extracted in the sample image I(x,y) (X,Y), then the feature Z extracted at this point is:
其中,w为滤波器窗口的大小，h(x,y,f0,θk,σx,σy)为去除直流分量后滤波器核心函数：Among them, w is the size of the filter window, h(x,y,f0 ,θk ,σx ,σy ) is the core function of the filter after removing the DC component:
其中f0与θk分别为波函数的频率和方向参数，σx和σy分别为高斯包络在x方向和y方向上的标准差。在I(x,y)抽取足够多的样点,则所有样点可以用由式Z＝(X,Y,f0,θk,σx,σy)提取特征。Where f0 and θk are the frequency and direction parameters of the wave function, respectively, and σx and σy are the standard deviations of the Gaussian envelope in the x and y directions, respectively. If enough samples are drawn in I(x, y), all samples can be extracted by the formula Z=(X, Y, f0 ,θk ,σx ,σy ).
计算各个通道滤波器后图像的均值和标准差作为特征组成特征向量，这里得到的特征向量值，仍然如图像上的像素点一样分布，我们按照特征值位置，经行特征值向量组化，得到n维特征值数据向量组。Calculate the mean and standard deviation of the image after each channel filter as the feature to form the eigenvector. The eigenvector values obtained here are still distributed like the pixels on the image. We group the eigenvalue vectors according to the position of the eigenvalues to get A set of n-dimensional eigenvalue data vectors.
(5)弹性网格特征值提取比对，这里主要是假想的网线对字体图像区域进行分割，这里是利用比例特性得出的字体图像外框，进行一定间隙的竖向划分和横向划分。图中若网格垂直方向和水平方向是均匀分布的，所以我们又称其为固定网格，若网格垂直方向和水平方向是非均匀分布的，我们又称其为弹性网格。在这里采用弹性网格，能容忍签名风格不同、局部自行变形等问题。我们利用归一化二值图像I4，当I4＝1时表示为黑像素点，I4＝0时表示为白像素点。然后将字体四方向分解后，我们提取横方向汉字子图像fH(x,y)，其中f(x,y)为汉字的二值图像，则横方向第i个网格内特性为：(5) Elastic grid eigenvalue extraction and comparison. Here, the imaginary network lines are used to divide the font image area. Here, the font image frame is obtained by using the proportional characteristics, and a certain gap is divided vertically and horizontally. In the figure, if the vertical and horizontal directions of the grid are evenly distributed, we call it a fixed grid, and if the vertical and horizontal directions of the grid are non-uniformly distributed, we call it an elastic grid. The elastic grid is used here, which can tolerate problems such as different signature styles and local self-deformation. We use the normalized binary image I4 , when I4 =1, it is represented as a black pixel, and when I4 =0, it is represented as a white pixel. Then, after decomposing the font in four directions, we extract the Chinese character sub-image fH (x, y) in the horizontal direction, where f(x, y) is the binary image of the Chinese character, then the internal characteristics of the i-th grid in the horizontal direction are:
其他几个方向的特征同理，“横、竖、撇、捺”四个方向特征值结合在一起，构成一个完整的字体特征矢量。这里得到的特征向量值，仍然如图像上的像素点一样分布，我们按照特征值位置，经行特征值向量组化，亦得到n维特征值数据向量组。The characteristics of the other directions are the same, and the eigenvalues of the four directions of "horizontal, vertical, left, and right" are combined to form a complete font feature vector. The eigenvector values obtained here are still distributed like the pixels on the image. We group the eigenvalue vectors according to the eigenvalue positions to obtain n-dimensional eigenvalue data vector groups.
(6)经过上述的比例特征、Gabor特征和弹性网格特征的提取，分别得到表示其特征的l维向量F1、d维向量F2及x维向量F3，其中F1＝{Fs1,Fs2,...,Fsl}，F2＝{Fg1,Fg2,...,Fgd}，F3＝{Fe1,Fe2,...,Fex}，然后分别对F1、F2和F3按照最大最小原则归一化特征向量为F1′、F′2和F′3即(6) After the above-mentioned extraction of proportional features, Gabor features and elastic grid features, l-dimensional vector F1 , d-dimensional vector F2 and x-dimensional vector F3 representing their features are respectively obtained, where F1 ={Fs1 ,Fs2 ,...,Fsl }, F2 ={Fg1 ,Fg2 ,...,Fgd }, F3 ={Fe1 ,Fe2 ,...,Fex }, and then respectively For F1 , F2 and F3 , according to the maximum and minimum principle, the normalized eigenvectors are F1 ′, F′2 and F′3 ie
最后，对归一化后的上述三种特征进行加权级联融合，即Finally, the weighted cascade fusion of the above three features after normalization is performed, namely
式中w1、w2和w3为经验权值，通过实验获取，且w1+w2+w3＝1。In the formula, w1 , w2 and w3 are empirical weights, obtained through experiments, and w1 +w2 +w3 =1.
(7)向量夹角余弦相似度测量，设由上述步骤所获得的待鉴定签名图像的融合特征向量可表示为F＝(x11,x12,...,x1n)，数据库中对应的签名图像的特征向量为G＝(x21,x22,...,x2n)。对于这两个n维向量，可以使用类似于夹角余弦的概念来衡量它们之间的相似程度，即：(7) Measurement of the cosine similarity between vector angles, assuming that the fused feature vector of the signature image to be identified obtained by the above steps can be expressed as F=(x11 ,x12 ,...,x1n ), the corresponding The feature vector of the signature image is G=(x21 , x22 , . . . , x2n ). For these two n-dimensional vectors, you can use a concept similar to the cosine of the angle to measure the similarity between them, namely:
夹角余弦取值范围为[-1,1]，夹角余弦越大表示两个向量的夹角越小，夹角余弦越小表示两向量的夹角越大。当两个向量的方向重合时夹角余弦取最大值1，当两个向量的方向完全相反夹角余弦取最小值-1。计算待鉴伪签名图像与数据库中对应的图像特征向量之间的夹角余弦值，将其与预先由实验所得出的经验阈值T作比较，若大于该阈值则鉴定为同一人的签名(反馈界面如图4所示)，若小于该阈值，则鉴定为别人伪造的签名(反馈界面如图5所示)。The value range of the included angle cosine is [-1,1]. The larger the included angle cosine, the smaller the included angle between the two vectors, and the smaller the included angle cosine, the larger the included angle between the two vectors. When the directions of the two vectors coincide, the cosine of the angle takes the maximum value of 1, and when the directions of the two vectors are completely opposite, the cosine of the angle takes the minimum value of -1. Calculate the cosine value of the included angle between the image of the false signature to be authenticated and the corresponding image feature vector in the database, compare it with the empirical threshold T obtained by the experiment in advance, if it is greater than the threshold, it will be identified as the signature of the same person (feedback interface as shown in Figure 4), if it is less than the threshold, it will be identified as someone else's forged signature (feedback interface as shown in Figure 5).
最后，我们利用现有的数据样本和仪器对系统的性能进行测试，测试结果如图7所示，基本完成了鉴定签字真伪的任务。Finally, we use the existing data samples and instruments to test the performance of the system. The test results are shown in Figure 7, basically completing the task of identifying the authenticity of the signature.
本发明可用其他的不违背本发明的精神和主要特征的具体形式来概括，因此，本发明的上述实施方案都只能认为是对本发明的说明而不能限制本发明，在与本发明的权利要求相当的含义和范围内任何改变，都应认为是包括在权利要求书的范围内。The present invention can be generalized by other specific forms that do not deviate from the spirit and main features of the present invention. Therefore, the above-mentioned embodiments of the present invention can only be regarded as explanations of the present invention and cannot limit the present invention. Any changes within the equivalent meaning and scope should be considered to be included in the scope of the claims.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410455357.9ACN104200239A (en) | 2014-09-09 | 2014-09-09 | Image feature fusion identification based signature authentic identification system and method |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410455357.9ACN104200239A (en) | 2014-09-09 | 2014-09-09 | Image feature fusion identification based signature authentic identification system and method |
| Publication Number | Publication Date |
|---|---|
| CN104200239Atrue CN104200239A (en) | 2014-12-10 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410455357.9APendingCN104200239A (en) | 2014-09-09 | 2014-09-09 | Image feature fusion identification based signature authentic identification system and method |
| Country | Link |
|---|---|
| CN (1) | CN104200239A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893952A (en)* | 2015-12-03 | 2016-08-24 | 无锡度维智慧城市科技股份有限公司 | Hand-written signature identifying method based on PCA method |
| CN106124508A (en)* | 2016-07-01 | 2016-11-16 | 宁夏六维辩证文物鉴定研究所 | A cultural relic identification system and identification method |
| CN106529395A (en)* | 2016-09-22 | 2017-03-22 | 文创智慧科技（武汉）有限公司 | Signature image recognition method based on deep brief network and k-means clustering |
| CN106599570A (en)* | 2016-12-12 | 2017-04-26 | 嘉兴太美医疗科技有限公司 | Informed consent checking method |
| CN107256389A (en)* | 2017-05-26 | 2017-10-17 | 山西农业大学 | The certification recognized based on handwriting characteristic and encryption method |
| CN107392136A (en)* | 2017-07-17 | 2017-11-24 | 河海大学常州校区 | A kind of feature based adaptively shakes the signature true or false identification system of decay |
| CN107798301A (en)* | 2017-10-13 | 2018-03-13 | 上海眼控科技股份有限公司 | A kind of signature detection system and method for vehicle annual test |
| CN108921126A (en)* | 2018-07-20 | 2018-11-30 | 北京开普云信息科技有限公司 | A kind of automatic identification signature stamp or the method and device of handwritten signature |
| CN109903247A (en)* | 2019-02-22 | 2019-06-18 | 西安工程大学 | High-precision grayscale method for color images based on Gaussian color space correlation |
| CN109902730A (en)* | 2019-02-21 | 2019-06-18 | 国网山东省电力公司临沂供电公司 | Detection method for broken strands in transmission lines based on deep learning |
| CN110007068A (en)* | 2019-03-25 | 2019-07-12 | 桂林优利特医疗电子有限公司 | A kind of urine drip detection method |
| CN110222660A (en)* | 2019-06-12 | 2019-09-10 | 河海大学常州校区 | A signature authentication method and system based on the fusion of dynamic and static features |
| CN110245615A (en)* | 2019-06-17 | 2019-09-17 | 河海大学常州校区 | A signature authentication method, system and storage medium |
| CN110245571A (en)* | 2019-05-20 | 2019-09-17 | 深圳壹账通智能科技有限公司 | Contract signature checking method, device, computer equipment and storage medium |
| CN110674771A (en)* | 2019-09-29 | 2020-01-10 | 北部湾大学 | Signature identification and verification method of Nixing pottery pottery |
| CN111062398A (en)* | 2020-03-18 | 2020-04-24 | 杭州汇萃智能科技有限公司 | Character recognition method, character recognition device, computer equipment and storage medium |
| CN111160245A (en)* | 2019-12-27 | 2020-05-15 | 上海眼控科技股份有限公司 | Dynamic signature identification method and device |
| CN111461015A (en)* | 2020-05-07 | 2020-07-28 | 中金金融认证中心有限公司 | RNN model-based user-independent online signature identification method and device |
| CN112102908A (en)* | 2020-09-22 | 2020-12-18 | 合肥易康达医疗卫生信息科技有限公司 | Credible cloud signature method for electronic medical record |
| CN112183348A (en)* | 2020-09-28 | 2021-01-05 | 中国建设银行股份有限公司 | Bill service processing method and device, electronic equipment and storage medium |
| WO2021031445A1 (en)* | 2019-08-22 | 2021-02-25 | 司法鉴定科学研究院 | Three-dimensional dynamic feature-based system and method for offline individual recognition by handwriting |
| CN112633148A (en)* | 2020-12-22 | 2021-04-09 | 杭州景联文科技有限公司 | Method and system for detecting authenticity of signature fingerprint |
| CN113034318A (en)* | 2021-05-06 | 2021-06-25 | 湖北经济学院 | River basin resource environment case data acquisition system and equipment |
| CN113139435A (en)* | 2021-03-30 | 2021-07-20 | 北京思特奇信息技术股份有限公司 | Self-learning signature handwriting deepening identification method and system |
| CN113313092A (en)* | 2021-07-29 | 2021-08-27 | 太平金融科技服务(上海)有限公司深圳分公司 | Handwritten signature recognition method, and claims settlement automation processing method, device and equipment |
| CN113468987A (en)* | 2021-06-17 | 2021-10-01 | 傲雄在线(重庆)科技有限公司 | Electronic handwriting authentication method, system, electronic equipment and storage medium |
| CN113723303A (en)* | 2021-08-31 | 2021-11-30 | 中国平安人寿保险股份有限公司 | Handwriting verification method and device, computer equipment and storage medium |
| CN114186205A (en)* | 2021-11-09 | 2022-03-15 | 海南火链科技有限公司 | Blockchain-based signature authentication method, device and computer equipment |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070154071A1 (en)* | 2005-12-29 | 2007-07-05 | Microsoft Corporation | Signature verification |
| CN101329735A (en)* | 2008-06-27 | 2008-12-24 | 北京中星微电子有限公司 | Device and method for identifying hand-written signature |
| KR101056307B1 (en)* | 2009-06-18 | 2011-08-11 | 인하대학교 산학협력단 | Multiline Reliability Support Hybrid Online Signature Verification System and Its Method for Various Applications |
| CN103927539A (en)* | 2014-03-24 | 2014-07-16 | 新疆大学 | Efficient feature extraction method for off-line recognition of Uyghur handwritten signature |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070154071A1 (en)* | 2005-12-29 | 2007-07-05 | Microsoft Corporation | Signature verification |
| CN101329735A (en)* | 2008-06-27 | 2008-12-24 | 北京中星微电子有限公司 | Device and method for identifying hand-written signature |
| KR101056307B1 (en)* | 2009-06-18 | 2011-08-11 | 인하대학교 산학협력단 | Multiline Reliability Support Hybrid Online Signature Verification System and Its Method for Various Applications |
| CN103927539A (en)* | 2014-03-24 | 2014-07-16 | 新疆大学 | Efficient feature extraction method for off-line recognition of Uyghur handwritten signature |
| Title |
|---|
| 牟小建: "基于Zernike矩的脱机手写签名鉴定研究", 《中国优秀博硕士学位论文全文数据库 (硕士) 信息科技辑》* |
| 高学等: "基于笔划的手写体汉字方向分解特征提取方法", 《华南理工大学学报（自然科学版）》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893952A (en)* | 2015-12-03 | 2016-08-24 | 无锡度维智慧城市科技股份有限公司 | Hand-written signature identifying method based on PCA method |
| CN106124508A (en)* | 2016-07-01 | 2016-11-16 | 宁夏六维辩证文物鉴定研究所 | A cultural relic identification system and identification method |
| CN106529395B (en)* | 2016-09-22 | 2019-07-12 | 文创智慧科技（武汉）有限公司 | Signature image identification method based on depth confidence network and k mean cluster |
| CN106529395A (en)* | 2016-09-22 | 2017-03-22 | 文创智慧科技（武汉）有限公司 | Signature image recognition method based on deep brief network and k-means clustering |
| CN106599570A (en)* | 2016-12-12 | 2017-04-26 | 嘉兴太美医疗科技有限公司 | Informed consent checking method |
| CN107256389A (en)* | 2017-05-26 | 2017-10-17 | 山西农业大学 | The certification recognized based on handwriting characteristic and encryption method |
| CN107392136A (en)* | 2017-07-17 | 2017-11-24 | 河海大学常州校区 | A kind of feature based adaptively shakes the signature true or false identification system of decay |
| CN107392136B (en)* | 2017-07-17 | 2020-07-17 | 河海大学常州校区 | A Signature Authenticity Identification System Based on Feature Adaptive Oscillation Attenuation |
| CN107798301A (en)* | 2017-10-13 | 2018-03-13 | 上海眼控科技股份有限公司 | A kind of signature detection system and method for vehicle annual test |
| CN108921126B (en)* | 2018-07-20 | 2019-07-23 | 北京开普云信息科技有限公司 | A kind of automatic identification signature stamp or the method and device of handwritten signature |
| CN108921126A (en)* | 2018-07-20 | 2018-11-30 | 北京开普云信息科技有限公司 | A kind of automatic identification signature stamp or the method and device of handwritten signature |
| CN109902730A (en)* | 2019-02-21 | 2019-06-18 | 国网山东省电力公司临沂供电公司 | Detection method for broken strands in transmission lines based on deep learning |
| CN109903247A (en)* | 2019-02-22 | 2019-06-18 | 西安工程大学 | High-precision grayscale method for color images based on Gaussian color space correlation |
| CN109903247B (en)* | 2019-02-22 | 2023-02-03 | 西安工程大学 | High-precision grayscale method of color image based on Gaussian color space correlation |
| CN110007068B (en)* | 2019-03-25 | 2022-07-08 | 桂林优利特医疗电子有限公司 | Urine leakage detection method |
| CN110007068A (en)* | 2019-03-25 | 2019-07-12 | 桂林优利特医疗电子有限公司 | A kind of urine drip detection method |
| CN110245571A (en)* | 2019-05-20 | 2019-09-17 | 深圳壹账通智能科技有限公司 | Contract signature checking method, device, computer equipment and storage medium |
| CN110222660A (en)* | 2019-06-12 | 2019-09-10 | 河海大学常州校区 | A signature authentication method and system based on the fusion of dynamic and static features |
| CN110222660B (en)* | 2019-06-12 | 2022-09-23 | 河海大学常州校区 | A signature authentication method and system based on the fusion of dynamic and static features |
| CN110245615A (en)* | 2019-06-17 | 2019-09-17 | 河海大学常州校区 | A signature authentication method, system and storage medium |
| CN110245615B (en)* | 2019-06-17 | 2022-09-13 | 河海大学常州校区 | Signature authentication method, system and storage medium |
| WO2021031445A1 (en)* | 2019-08-22 | 2021-02-25 | 司法鉴定科学研究院 | Three-dimensional dynamic feature-based system and method for offline individual recognition by handwriting |
| CN110674771A (en)* | 2019-09-29 | 2020-01-10 | 北部湾大学 | Signature identification and verification method of Nixing pottery pottery |
| CN110674771B (en)* | 2019-09-29 | 2022-04-01 | 北部湾大学 | Nixing pottery ware signature identification and verification method |
| CN111160245A (en)* | 2019-12-27 | 2020-05-15 | 上海眼控科技股份有限公司 | Dynamic signature identification method and device |
| CN111062398A (en)* | 2020-03-18 | 2020-04-24 | 杭州汇萃智能科技有限公司 | Character recognition method, character recognition device, computer equipment and storage medium |
| CN111461015A (en)* | 2020-05-07 | 2020-07-28 | 中金金融认证中心有限公司 | RNN model-based user-independent online signature identification method and device |
| CN111461015B (en)* | 2020-05-07 | 2023-10-13 | 中金金融认证中心有限公司 | RNN (RNN) -model-based user-independent online signature identification method and device |
| CN112102908A (en)* | 2020-09-22 | 2020-12-18 | 合肥易康达医疗卫生信息科技有限公司 | Credible cloud signature method for electronic medical record |
| CN112183348B (en)* | 2020-09-28 | 2024-10-15 | 中国建设银行股份有限公司 | Bill service processing method and device, electronic equipment and storage medium |
| CN112183348A (en)* | 2020-09-28 | 2021-01-05 | 中国建设银行股份有限公司 | Bill service processing method and device, electronic equipment and storage medium |
| CN112633148B (en)* | 2020-12-22 | 2022-08-09 | 杭州景联文科技有限公司 | Method and system for detecting authenticity of signature fingerprint |
| CN112633148A (en)* | 2020-12-22 | 2021-04-09 | 杭州景联文科技有限公司 | Method and system for detecting authenticity of signature fingerprint |
| CN113139435A (en)* | 2021-03-30 | 2021-07-20 | 北京思特奇信息技术股份有限公司 | Self-learning signature handwriting deepening identification method and system |
| CN113034318A (en)* | 2021-05-06 | 2021-06-25 | 湖北经济学院 | River basin resource environment case data acquisition system and equipment |
| CN113468987A (en)* | 2021-06-17 | 2021-10-01 | 傲雄在线(重庆)科技有限公司 | Electronic handwriting authentication method, system, electronic equipment and storage medium |
| CN113313092A (en)* | 2021-07-29 | 2021-08-27 | 太平金融科技服务(上海)有限公司深圳分公司 | Handwritten signature recognition method, and claims settlement automation processing method, device and equipment |
| CN113313092B (en)* | 2021-07-29 | 2021-10-29 | 太平金融科技服务(上海)有限公司深圳分公司 | Handwritten signature recognition method, and claims settlement automation processing method, device and equipment |
| CN113723303A (en)* | 2021-08-31 | 2021-11-30 | 中国平安人寿保险股份有限公司 | Handwriting verification method and device, computer equipment and storage medium |
| CN113723303B (en)* | 2021-08-31 | 2025-03-18 | 中国平安人寿保险股份有限公司 | A handwriting verification method, device, computer equipment and storage medium |
| CN114186205A (en)* | 2021-11-09 | 2022-03-15 | 海南火链科技有限公司 | Blockchain-based signature authentication method, device and computer equipment |
| Publication | Publication Date | Title |
|---|---|---|
| CN104200239A (en) | Image feature fusion identification based signature authentic identification system and method | |
| CN106778586B (en) | Off-line handwritten signature identification method and system | |
| CN111401372B (en) | A method for extracting and identifying image and text information from scanned documents | |
| CN110222660B (en) | A signature authentication method and system based on the fusion of dynamic and static features | |
| Bana et al. | Fingerprint recognition using image segmentation | |
| Wu et al. | Palm line extraction and matching for personal authentication | |
| CN103324944B (en) | A False Fingerprint Detection Method Based on SVM and Sparse Representation | |
| CN104809451B (en) | A kind of person's handwriting identification system based on stroke curvature measuring | |
| US7724958B2 (en) | Systems and methods for biometric identification using handwriting recognition | |
| CN100395770C (en) | A Hand Feature Fusion Authentication Method Based on Feature Relationship Measurement | |
| CN101957919B (en) | Character recognition method based on image local feature retrieval | |
| CN104123537A (en) | Rapid authentication method based on handshape and palmprint recognition | |
| CN102332084B (en) | Identity identification method based on palm print and human face feature extraction | |
| CN110008909A (en) | An AI-based real-name business real-time audit system | |
| CN104464079A (en) | Multi-currency-type and face value recognition method based on template feature points and topological structures of template feature points | |
| CN102156884A (en) | Straight segment detecting and extracting method | |
| CN104680154B (en) | An identity recognition method based on the fusion of face features and palmprint features | |
| CN106874968B (en) | Second-generation identity card authenticity identification method | |
| CN103927539A (en) | Efficient feature extraction method for off-line recognition of Uyghur handwritten signature | |
| CN116778503A (en) | Seal authenticity identification method and system based on OCR (optical character recognition) | |
| Su et al. | Automatic seal imprint verification systems using edge difference | |
| Chen et al. | Fingerprint matching with registration pattern inspection | |
| Su et al. | Svm-based seal imprint verification using edge difference | |
| Dewangan | Real Time Recognition of Handwritten Devnagari Signatures without Segmentation Using Artificial Neural Network | |
| Kumar et al. | Inter-point envelope based distance moments for offline signature verification |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20141210 | |
| RJ01 | Rejection of invention patent application after publication |