• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    This patent proposes a method and system to generate manuscript text with different degrees of maturity of the writer without having to have the writer's original text. Additionally, by generating both online text (synthesizing the temporal movement) and offline (generating the manuscript image), it allows to simulate the real evolution both in the number of strokes and in the speed profile of the writer according to his progress. It makes it possible to generate writings of writers with possible neurological and motor dysfunctions. (Machine-translation by Google Translate, not legally binding)
    公开号:ES2610794A1
    申请号:ES201600539
    申请日:2016-06-17
    公开日:2017-05-03
    发明作者:Miguel Ángel FERRER BALLESTER；María Cristina CARMONA DUARTE
    申请人:Universidad de las Palmas de Gran Canaria；
    IPC主号:
    专利说明:

    DESCRIPTION

    Method and system to generate handwritten text with different degrees of maturity of the writer.
     5
    The present invention relates to the generation of manuscript text, more particularly to a method and system for generating manuscript text from synthetic writers and its evolution over time both for online text (synthesizing temporal movement) and offline (generating the image of the manuscript ).
     10
    Field of the invention

    The invention falls within the technical sector of handwriting text synthesis. The invention relates to methods for generating handwritten text by means of a computing device as similar as possible to human text. fifteen

    The methods present in the current state of the art do not solve the problem of generating handwritten text with different degrees of maturity of the writer, that simulates the evolution of the writer and that allows, by the minimum number of parameters, to simulate both the positions of the pen or pencil in time as the mode of writing 20 depending on the age or learning time of the writer.

    This patent proposes a method and system to generate handwritten text with different degrees of maturity of the writer without having to have original text. It allows to simulate the real evolution both in number of strokes and in the speed profile of the writer according to his progress. It allows to generate manuscript of writers with possible neurological and motor dysfunctions.

    This new method and system allows to solve technical problems in the OCR of handwritten text, improves the identification of the writer, allows a follow-up of the learning 30 of writing in children and provides an improvement in the monitoring and screening of neurodegenerative diseases, as well as dysgraphia.

    Background of the invention
     35
    Research on manuscript synthesis is used in different fields. A first motivation was to provide a large database of biometric writing to evaluate the verifiers of automated signatures or automated identifiers of the writer and avoid legal problems about privacy [US 7227993 (B2). Today it is understood that inclusion in a manuscript synthesizer of intra and inter person variability and health variability due to different diseases such as Parkinson's or Alzheimer's, etc. They could help improve understanding of the underlying processes in the production of human manuscript. Artistic creations and captcha generation are another motivation [Chetan Ramaiah, Rèjean Plamondonm, and Venu Govindaraju, A Sigma-Lognormal Model for Handwritten Text CAPTCHA Generation, in: Proceedings of the 45 International Conference on Pattern Recognition. 2014. pp. 250-255], [O. Thomas A. Rusu, V. Govindaraju, Synthetic handwritten CAPTCHAs. Pattern Recognition 42 (2009) 3365-3373].

    Some patents try to solve the technical problem of generating synthetic handwritten text. For example, US 5327342 (A) refers to providing a method for producing simulated handwriting and, more particularly, an apparatus and a method for producing handwriting that simulates the user's own handwriting.
    WO 2006042307 (A2) refers to providing a method and apparatus for generating a realistic handwritten communication.

    Patent document US 7227993 (B2) refers to providing a system based on learning and the process of synthesis of cursive writing. 5

    Patent document US 7983478 B2 refers to the generation of calligraphy letters based on a Markov model procedure.

    Patent document G8 2394348 refers to a method to synthesize from a user's handwriting data, by dividing into glyphs and storing them in a memory, it generates other letters of the same user.

    Patent document US 7227993 B2 presents a method that generates cursive letters similar to those of the user. It is also based on taking an image of the letter of the user (glyphs) and adding variations in the image. It does not simulate the speed profile.

    These methods are based on processing the image of the handwritten text, which increases the processing time. Nor do they simulate the learning and variability obtained with practice or give the pencil positions with their temporal reference. twenty

    Patent document WO 2000078215 (A1) refers to a method for the analysis of motor ability based on the measurement and monitoring of motor control by approximating a sum of lognormal with different parameters of the velocity profile obtained from writing. On a digitizer tablet. 25

    WO 1992014219 (A1) refers to a method for digitizing and segmenting the movement of writing based on curvilinear motion and angular velocity. From a previous movement of the user, it parameterizes the data to recover it again. 30

    WO 2009018983 (A1) refers to a method that encourages writing using a virtual pen that writes and draws synthesizing different types of forms and texts.
     35
    In turn, the work that is closest to solving the technical problem of generating signatures with synthetic manuscript text giving the possibility of generating variability in the signatures and in the text has been presented in [Ferrer. M.A .; Diaz-Cabrera, M .; Morales, A., "Static Signature Synthesis: A Neuromotor lnspired Approach for Biometrics", Pattern Analysis and Machine Intelligence, IEEEE Transactions on, vol. PP, no.99, pp.1, 1 (2015)]. 40 The method is based on fixing the position and trajectory of the character to be generated in the grid points, all the characters generated are joined to form words and then a method is applied to generate the text as if it were real by setting the values of some filters that simulate the neuromuscular model. This model generates variability by varying the distances in the grid, the inclination of the grid and varying the parameters of the filters to simulate letters written at different speeds. The disadvantage of this model is that it cannot simulate the variability that occurs naturally in a writer over time and does not allow varying the parameters to simulate writers of different ages.

    These previous methods do not solve the problem of generating written text without using 50 real written text, which simulates the different ages of the writer and allows, by the minimum number of parameters, to simulate both the positions of the pen or pencil in time and the way of writing depending on the age, degree of maturity or learning time of the writer.

    Therefore, they remain as unsolved technical problems: synthesize text similar to the real one, simulating the real evolution both in number of strokes and in the speed profile of the writer according to his progress and giving the possibility of varying parameters that can be related to neurological parameters or Human learning engines in a simple way.

    Summary of the invention

    The present invention relates to a new method and system for generating manuscript text 10 valid for writers with different degrees of maturity.

    The method of the present invention, unlike other solutions proposed at international level, allows to generate valid manuscript text for writers with different degrees of maturity without having to have the original text of the previous writer, simulating the real evolution both in numbers of strokes and in the writer's speed profile according to his progress and giving the possibility to vary parameters that can be related to neurological parameters or motors of human learning in a simple way.

    The invention is characterized by the following steps:

    (i) for each letter that is intended to be synthesized, its path vectors that have been previously stored in a memory, consisting of an nsl number of points, are selected,
     25
    (ii) the method for generating the temporal evolution in manuscript text of claim 2 is applied, selecting the degree of maturity of the writer by parameter E,

    (iii) connect each of the letters that make up the word, 30

    (iv) the word o / and the velocity profile are synthesized by applying the kinematic adapted method. The characteristics of the synthetic writer are established by defining the parameters (Kμ, Kσ, Kt, KD, Kα, εD, εt).
     35
    And the method to generate the temporal evolution in handwritten text that is characterized by selecting a percentage E of the initial points of which the trajectory of each letter that has been previously stored in a memory is recorded. The percentage E is inversely proportional to the evolution or maturity of the writing.
     40
    The details of one or more embodiments of what is described in this summary are set forth in the following accompanying drawings and description. Other features, aspects and advantages of the subject will be apparent from the description, the drawings. and the claims.
     Four. Five
    Brief description of the figures

    These are presented by way of illustration and not limitation of the invention. The following figures show:
     fifty
    • Figure 1 a diagram describing the method according to the invention.

    • Figure 2 a diagram describing the method of evolution.

    • Figure 3 an image with an example of the text generated with the method.

    • Figure 4 an example of synthesized letters of different ages and the corresponding speed of the writer compared to real text.
     5
    • Figure 5 an example of an adult's velocity profile obtained from drawing an intermediate figure arc) with an angle less than 50 °. It is observed in the upper figure that the path is formed by two vectors, giving rise to a velocity profile with two independent signals (lower figure, dashed line) that model the real velocity profile (lower figure, continuous line). 10

    Detailed description of the invention

    Although the invention is described in terms of a specific preferred embodiment, it will be apparent to those skilled in this art that various modifications, redispositions and replacements can be made. The scope of the invention is defined by the claims appended thereto.

    The present invention allows to solve the technical problem of synthesizing text similar to the real one without having to have the original text of the previous writer, simulating the real evolution in both the number of strokes and the speed profile of the writer according to his progress and giving the possibility of vary parameters that can be related to neurological parameters or motors of human learning in a simple way.

    Detailed description of a preferred embodiment of the invention

    The present invention can be implemented in hardware, software or a combination of both.

    Figure 1 shows the different stages of the procedure and how it is possible to solve the technical problem. In the first stage (101) it consists in selecting the trajectory of each letter starting from a previously stored trajectory (103), a process that is performed only once when the type of writing or language to be learned is programmed. Next, the evolution method (105) is applied, which selects a series of points depending on the age of the writer (Figure 2) and is explained in the following subsection. The different letters are then joined, making sure that the upper (201) and lower lines of the grid (204) are joined.

    In the proposed method it is considered unlike that in [Ferrer. M.A .; Diaz-Cabrera, M .; Morales, A., "Static Signature Synthesis: A Neuromotor lnspired Approach for Biometrics", 40 Pattern Analysis and Machine Intelligence. IEEE Transactions on, vol. PP, no. 99. pp. 1.1 (2015)] that every two consecutive grid points is considered a vector (Figure 2) and also for each vector the angle αj between the previous and the current vector is also defined, which is defined as:
     Four. Five


    With these previous calculations it is considered that for each vector the parameters Dj, t0j, μj and σj of a signal that models the velocity profile are adjusted and that
    recommends that it be a lognormal (or another that adjusting its parameters of the same setting):


     5
    where μ and σ are the location and scale parameters respectively, j indicates the number of the corresponding vector and t0j is the initial moment where the beginning of the lognormal is located.

    The following explains the equations and premises that these 10 parameters have to have in order for the manuscript text to be synthesized with real appearance and to include the requirements for the writer's evolution to be similar to the real one.

    In this method the value σj, which defines the width of each lognormal, is defined for each
    lognormal similar and given by equation: 15


    where Kσ is a definite constant for each writer and different for different writers. To generate evolution, the parameter Kσ approaches 0.01 to simulate writing of children and it is varied up to a value of 0.05 for adults to obtain about 20 results as in Figure 3. If you want to approximate that of a particular person this value it would be adjusted with the last lognormal of the speed profile of the real person (Figure 5), when the person makes a quick curved stroke of angle less than 60º of the size of the text that is to be synthesized.
     25
    The initial period of each lognormal t0j is expressed as:



    where nsl is the number of points that form the trajectory of the letter, Kt is a constant that can be set for each writer, to set the small scales of the letters and the average speed, Kα is the maximum delay time due to the angle, N (0, εt is a random variable that emulates the dispersion and is inversely proportional to writing ability and finally the delay introduced due to the angle between the vectors (αj) is given by a sigmoid curve (S (α, 1 / 15, - 70)), where: 35



    Kt = 0.04, Kα = 0.1 and εt = +/- 0.02 are set as the initial setting. These values can be adjusted to approximate them to the real parameters of a writer as explained with the angle. In this case, the parameter Kα would be adjusted with the greater increase in time (Δt) (Figure 5 below) in the velocity profile of that of the real person, when the person makes a curved rapid stroke of angle less than 5º of the text size that you want to synthesize.
     Four. Five
    The amplitude D¡ of the lognormal j is defined proportional to the distance between the points of the path:


     5
    Where dact, j is the distance between the points of the path, Q¡ to Qj-1, dref is the minimum distance between the grid points., KD is a constant dependent on each individual and can also be used for large scaling of the letters in the cases for which Kt is not enough. N (0, εD) is a random value for each lognormal j. The initial setting is KD = dref, j, and εD = 0.3. These values can be adjusted using an approximation 10 to adjust them to the real parameters of a writer, when this is his end.

    The variable μj is equal to a constant Kμ, a constant that depends on the motor characteristics of the subject and therefore is different for each individual. In the experiment performed to obtain Figure 3 it was set to zero. fifteen

    image 1
    Once all the parameters for each component of the path (t; to¡) have been calculated, it is obtained by substituting the parameters Dj, t0j, μj and σj in Equation 2. Finally, the final velocity vector is obtained by the sum of each of the (t; toj) based on [M Djioua. R. Plamondon. A new algorithm and system for the characterization of handwriting 20 strokes with delta-lognormal parameters. IEEEE Transactions on Pattern Analysis and Machina Intelligence, 31 (2009) 2060-2072]:
    image 1


     25
    image 1
    The x and y components of the vector (t) are adjusted by calculating the angle of each path Øj.


     30
    image 1
    And replacing it in the equations to calculate vx, j (t; toj) and vy, j (t; toj) from (t; toj) using (Equation 9) and (Equation 10).


     35



    Finally, the Cartesian path that is used to draw the letter is obtained:
     40




     Four. Five
    In order to synthesize problems in the learning or neurological health of the individual, the parameters εD and εt are modified, increasing its value the more separated from the normal value is the ideal individual to be synthesized.
    To synthesize patients with Parkinson's, the increase in time between the signals representing each vector (Kt) is increased. In this way, the more affected by the disease, the smaller the overlap between the signals that model each of the vectors. The vectors that make up the trajectory are designed in pairs, and in each of the pairs the direction of one counteracts the previous one (simulates the agonist and antagonistic movement of the muscles).

    To synthesize Alzheimer's patients, the value of the parameters related to the starting point stability of the signal representing each vector and the path error (εD, εt) is increased. The higher the value of these parameters, the greater the degree of degeneration due to the aforementioned disease in adults.

    To synthesize patients with amyotrophic lateral sclerosis, the width of the signal representing each path vector is increased by varying Kσ or Kμ and decreasing the amplitude (KD). The greater the width and the smaller the amplitude the greater the degree of degeneration due to the disease in adults.

    Alternative method
     twenty
    Instead of using Equation 1, a sinusoidal can be used, where:



    The procedure would be the same as in the previous method except that in this case the value of 25 σj (Equation 3) would be multiplied by KDs, as follows:



    The value of KDs depends on the writer as does Ks. To simulate a real writer, the value is adjusted by comparing the value synthesized with the rest of the parameters and KD equal to 1 with the speed signal obtained from a real sample. The real sample will be obtained from the one captured using a digitizing tablet on which the real writer will draw a 1 cm straight line.
     35
    Method for temporal evolution

    This section details the block (103) and shown in detail in (Figure 2).

    It is based on this method that as the number of points used in the writing is learned 40 is smaller and therefore the dact length, j is greater. For this, a variable E is defined that defines the percentage of points selected for each individual. Therefore the number of outputs that the system will have will be given by L:

     Four. Five

    This method is detailed in the figure (Figure 2) and consists of:

    1. Those of nsl - 3 of the path points are segmented into L clusters. Each cluster distances itself in (nsl - 3) / (L - 2). For the first and the last, the maximum size is taken into account to preserve readability.

    2. One of the grid points is randomly selected from each cluster. 5

    3. To ensure readability, at least we have to preserve, if any, one of the points that matches any of the lines set for each language (201) (202) (203) (204). In the event that the rule is not met, return to point e until the readability standard is met. 10

    Once the grid points have been calculated, the parameters are calculated using the previous method.

    Examples
     fifteen
    The invention will now be illustrated by tests carried out by the inventors, which show the effectiveness of the invention.

    These specific examples provided serve to illustrate the nature of the present invention and are included for illustrative purposes only, and therefore should not be construed as limitations on the invention claimed herein. Therefore, the examples described below illustrate the invention without limiting its scope of application.

    Example 1. Synthesized text automatically offline. 25

    An example of the different manuscripts generated from a single initial grid configuration and path by letter is shown in Figure 3. These results have been obtained by varying the evolution parameter E from E = 100 to synthesize the writing of a child (first row of text) to E = 20 to synthesize the writing of an adult person who writes frequently. In addition, to improve its realism, the value of εD and εt has varied linearly from εD = 0.3 and εt = 0.02 when E = 100 to 0 when E = 20. You can see how in the text synthesized for E = 20 the letters have more variability than for those generated with E = 100. This is because in the case of E = 20 random path points are selected which makes the number of possible 35 variations much higher than in a child. Therefore the parameter E, according to its value is lower, also synthesizes different writers while when E = 100 a single writer or individual is being synthesized, with its own variability.

    Example 1. Text automatically synthesized online. 40

    An example of the letter 'a' real and synthetically generated both online and offline and its speed profiles is shown in Figure 4.

    To obtain the real samples, 20 children (aged 5 and 10) from the same school were asked to write the letter "a" on a Wacom digitizer tablet. The digitizing tablet is a WACOM Intuos 3 with an "Intuos 3 Grip Pen" pen, with a sampling frequency of 200 Hz and a resolution of 2540 dpi and a working surface of 304.8 mm x 228.6 mm. The same process was carried out with 10 adults over 18 years. fifty

    In figure 4, the left column shows the real results obtained from one of the cases for each age, having been verified that the other cases of the same age had the same characteristics. It can be seen as in the case of
    5-year-old children appear many more spikes in the speed profile than in the adult.

    In Figure 4, the results obtained with the letter "a" synthesized by the method proposed in this patent are shown in the right column. 5 can be seen as in the velocity profile more peaks also appear when child letter is synthesized with a value of E = 100, instead only three peaks appear in the case of synthesizing adult letter using a value of E = 20, equal What happens for real samples.
    权利要求:
    Claims (13)
    [1]

    1. Method and system to generate valid manuscript text for writers with different degrees of maturity characterized by the following stages:
     5
    (i) for each letter that is intended to be synthesized, its path vectors that have been previously stored in a memory, consisting of an nsl number of points, are selected,
    (ii) the method for generating the temporal evolution in manuscript text of claim 2 is applied, selecting the degree of maturity of the writer by means of parameter E,
    (iii) connect each of the letters that make up the word,
     fifteen
    (iv) the word o / and the velocity profile are synthesized by applying the kinematic adapted method. The characteristics of the writer synthesize are established by defining those for meters (Kμ, Kσ, Kt, KD, Kα, εD, εt).

    [2]
    2. Method to generate the temporal evolution in handwritten text that is characterized by selecting a percentage E of the initial points of which the trajectory of each letter that has been previously stored in a memory consists. The percentage E is inversely proportional to the evolution or maturity of the writing.

    [3]
    3. Method according to claim 1 and 2 characterized by being online and generating the 25-speed profile after point (vi) taking into account the degree of maturity of the writer.

    [4]
    4. Method according to claim 1 and 2 characterized by being an offline method and applying an ink model after point (vi) (202).
     30
    [5]
    5. Method according to claim 1 and 2 characterized by obtaining the writing characteristics of a real person by adjusting the parameters Kμ, Kt, Kσ, Kt, Kα and KD from the real speed profile. In addition, the path vectors that have been previously stored in a memory that are used in point (i) are obtained from the actual writing of the writer to be synthesized. 35

    [6]
    6. Method according to claim 1 characterized in that the points (ii) and (iii) can be exchanged.

    [7]
    7. Method according to claims 1 and 2 characterized by defining the value of E from 40 100 to 20 to emulate the evolution from child to adult

    [8]
    8. Method according to claims 1 and 2 characterized in that the parameters εD and εt are related to the learning or neurological health of the individual. The higher the value of these parameters, the worse your neurological health or worse the learning. Four. Five

    [9]
    9. Method according to claims 1 and 2 characterized by synthesizing patients with Parkinson's increasing the time increase between the signals representing each vector (Kt). In this way, the more affected by the disease, the lower the patient is the overlap between the serials that model each of the vectors. It is considered that the vectors that make up the trajectory are designed in pairs, and in each of the pairs the direction of one counteracts the previous one (simulates the agonist and antagonistic movement of the muscles).
    [10]
    10. Method according to claims 1 and 2 characterized by synthesizing Alzheimer's patients by increasing the parameters related to the starting point stability of the signal representing each vector and the path error (εD, εt). The higher the value of these parameters, the greater the degree of degeneration due to the aforementioned disease in adults. 5

    [11]
    11. Method according to claims 1 and 2 characterized by synthesizing patients with amyotrophic lateral sclerosis by increasing the width of the signal representing each path vector (Kσ or Kμ) and decreasing its amplitude (KD). The larger the width and the smaller the amplitude the greater the degree of degeneration 10 due to the disease in adults.

    [12]
    12. Electronic device comprising a screen, at least one processor, at least one memory and at least one computer application stored in said memory and configured to be executed by the processor and wherein said application 15 includes instructions for generating handwritten text such as claim 1 .

    [13]
    13. Computer program with instructions configured for execution by at least one processor which, when executed by an electronic device, causes it to carry out the procedure of any of the preceding claims. twenty
    类似技术:
    公开号 | 公开日 | 专利标题
    Ferrer et al.2014|Static signature synthesis: A neuromotor inspired approach for biometrics
    Pizlo et al.2014|Making a machine that sees like us
    Kristensson et al.2012|Continuous recognition of one-handed and two-handed gestures using 3D full-body motion tracking sensors
    Hoffman et al.2010|Breaking the status quo: Improving 3D gesture recognition with spatially convenient input devices
    BRPI0721562A2|2013-01-22|Volume Recognition Method and System
    CN103020085B|2015-09-30|A kind of implementation method of personal style calligraphy dynamic word-base and application thereof
    Mohan et al.2011|Teaching a humanoid robot to draw ‘Shapes’
    Vanderloock et al.2013|The skweezee system: enabling the design and the programming of squeeze interactions
    Kupin et al.2019|Task-driven biometric authentication of users in virtual reality | environments
    CN104345948A|2015-02-11|Implementation method and implementation device for original handwriting as well as electronic equipment
    Yeh et al.2002|On the effects of haptic display in brush and ink simulation for chinese painting and calligraphy
    ES2610794B2|2017-09-19|Method and system to generate handwritten text with different degrees of maturity of the writer
    Shaw et al.2016|Analyzing the articulation features of children's touchscreen gestures
    Bodenheimer et al.1999|The effects of noise on the perception of animated human running
    CN103823628A|2014-05-28|Calligraphy stoke simulating method combined with pen pressure
    Lee et al.1998|A new synthesizing method for handwriting Korean scripts
    André et al.2014|A parsimonious oscillatory model of handwriting
    Khan et al.2017|Use hand gesture to write in air recognize with computer vision
    CN107122042A|2017-09-01|The Chinese-character writing method and system that a kind of quiet dynamic gesture is combined
    Singla et al.2019|Visual rendering of shapes on 2D display devices guided by hand gestures
    JP3802483B2|2006-07-26|Calligraphy learning support system, computer, program, and recording medium
    Lv et al.2017|Smart motion reconstruction system for golf swing: a DBN model based transportable, non-intrusive and inexpensive golf swing capture and reconstruction system
    Plamondon et al.2020|The Lognormality Principle and its Applications in e-security, e-learning and e-health
    Barneva et al.2016|Motion tracking for gesture analysis in sports
    Plamondon2021|The lognormality principle: a personalized survey
    同族专利:
    公开号 | 公开日
    ES2610794A9|2017-08-10|
    ES2610794B2|2017-09-19|
    US10977422B2|2021-04-13|
    WO2017216400A1|2017-12-21|
    US20190303422A1|2019-10-03|
    CA3028063A1|2017-12-21|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US20040148577A1|2003-01-27|2004-07-29|Ying-Qing Xu|Learning-based system and process for synthesizing cursive handwriting|
    US20120172682A1|2005-12-21|2012-07-05|Norconnect Inc.|Method and apparatus for biometric analysis using eeg and emg signals|
    US5077802A|1991-02-11|1991-12-31|Ecole Polytechnique|Apparatus and method for digitizing and segmenting a handwriting movement based on curvilinear and angular velocities|
    US5327342A|1991-03-31|1994-07-05|Roy Prannoy L|Method and apparatus for generating personalized handwriting|
    CA2275733A1|1999-06-16|2000-12-16|Wacef Guerfali|Method and device for measuring, analysing and monitoring motor aptitudes and performances for training, rehabilitation and detection of improper functioning|
    GB2394348B|2002-10-16|2005-09-28|Canon Kk|Handwriting synthesis|
    US7352899B2|2004-10-12|2008-04-01|Loeb Enterprises, Llc|Realistic machine-generated handwriting with personalized fonts|
    ITPI20070093A1|2007-08-08|2009-02-09|Mario Pirchio|METHOD TO ANIMATE ON THE SCREEN OF A COMPUTER A PENNAVIRTUAL WRITING AND DRAWING|
    US7983478B2|2007-08-10|2011-07-19|Microsoft Corporation|Hidden markov model based handwriting/calligraphy generation|
    US10055659B2|2015-05-04|2018-08-21|King Fahd University Of Petroleum And Minerals|Systems and associated methods for arabic handwriting using synthetic extension concatenation and synthesis|
    US10140262B2|2015-05-04|2018-11-27|King Fahd University Of Petroleum And Minerals|Systems and associated methods for Arabic handwriting synthesis and dataset design|WO2021097573A1|2019-11-21|2021-05-27|Polyvalor, Limited Partnership|Method and system for assessing a biological movement|
    法律状态:
    2017-09-19| FG2A| Definitive protection|Ref document number: 2610794 Country of ref document: ES Kind code of ref document: B2 Effective date: 20170919 |
    优先权:
    申请号 | 申请日 | 专利标题
    ES201600539A|ES2610794B2|2016-06-17|2016-06-17|Method and system to generate handwritten text with different degrees of maturity of the writer|ES201600539A| ES2610794B2|2016-06-17|2016-06-17|Method and system to generate handwritten text with different degrees of maturity of the writer|
    US16/308,305| US10977422B2|2016-06-17|2017-06-01|Method and system for generating handwritten text with different degrees of maturity of the writer|
    PCT/ES2017/000070| WO2017216400A1|2016-06-17|2017-06-01|Method and system for generating handwritten text with different degrees of maturity of the writer|
    CA3028063A| CA3028063A1|2016-06-17|2017-06-01|Method and system for generating handwritten text with different degrees of maturity of the writer|
    [返回顶部]