DATA PROCESSING SYSTEMS AND METHODS FOR ASSESSING AN INDIVIDUAL'S FINE MOTOR SKILLS

Abstract

One of the objectives of this invention is to provide a solution that allows for an objective and precise assessment of graphomotor skills. To this end, the inventors propose to evaluate, separately, the different facets of graphomotor skills produced during an activity performed on an interactive manipulation surface. In practice, each facet of the graphomotor skill is evaluated by a respective evaluation model. Then, for a current activity, which is evaluated by a current evaluation model, a machine learning model is used that has been trained to predict the next evaluation model to use, based on the results obtained by the current evaluation model. In this way, professionals involved in education and health will benefit from a reliable method for assessing fine motor skills without being influenced by their own subjectivity. Figure to be published with the abbreviation: Figure 1

Description

Title of the invention: DATA PROCESSING SYSTEMS AND METHODS FOR ASSESSING AN INDIVIDUAL'S FINE MOTOR SKILLS technical field

[0001] The invention relates to the field of gesture recognition using a computer tool. In particular, it relates to systems and methods for data processing for evaluating an individual's fine motor skills using an interactive manipulation surface. Previous technique

[0002] The evaluation of graphomotor gesture is of crucial importance in determining the fine motor skills and coordination of an individual, particularly in developing children.

[0003] Generally, this assessment is carried out subjectively by professionals such as teachers, occupational therapists or psychomotor therapists who base it on direct observation and their own experiences.

[0004] The subjectivity of these evaluations stems from the fact that different people can potentially interpret the same graphomotor gesture differently.

[0005] Evaluation criteria are not always clearly defined and may vary depending on the evaluator's training, personal skills and professional preferences.

[0006] Moreover, the lack of reliable quantitative tools to measure the quality of the gesture makes it difficult to make an objective comparison between different individuals or evaluation sessions.

[0007] This can lead to errors in the assessment and create some uncertainty as to the individual's actual performance with regard to their graphomotor skills.

[0008] However, such a situation is problematic, as it can lead to an erroneous diagnosis, inappropriate care or insufficient monitoring of progress.

[0009] Thus, there is a need for a solution that allows an objective and precise evaluation of the graphomotor gesture. Summary of the invention

[0010] The invention aims to solve, at least partially, this need.

[0011] A first aspect of the invention relates to a data processing system for evaluating an individual's fine motor skills in the context of performing a routine activity that involves graphomotor gestures in carrying out at least one drawn on an interactive manipulation surface.

[0012] In particular, the evaluation is carried out on the basis of a single current evaluation model which is chosen from a plurality of evaluation models.

[0013] In practice, each evaluation model describes, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators which are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index which is representative of a combination of at least two kinematic and / or kinetic indicators.

[0014] The data processing system includes at least one sensor which is coupled to the interactive manipulation surface.

[0015] In practice, the sensor is designed to detect and record at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when carrying out at least one tracing within the context of the activity.

[0016] In particular, the sequence of movements comprises a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes.

[0017] The data processing system also includes at least one processor which is coupled to the sensor.

[0018] In practice, the processor is designed to calculate a plurality of kinematic and / or kinetic indicators from the attributes that are associated with all or part of the plurality of sequential trajectory points of the sequence of movements,

[0019] Furthermore, the processor is also designed to execute the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index,

[0020] Then, the processor is also designed to calculate at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model.

[0021] Next, the processor is also designed to form a vector which includes, a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index, and the calculated gap.

[0022] Furthermore, the processor is also designed to provide the input vector for a machine learning model trained to predict the next activity evaluation model.

[0023] Finally, the processor is also designed to execute the trained machine learning model in such a way as to obtain, as output from the trained machine learning model, a third piece of information that is representative of the next evaluation model to be used to assess the activity.

[0024] In a first embodiment of the first aspect of the invention, the predetermined threshold includes a success threshold and / or a failure threshold.

[0025] In a second embodiment of the first aspect of the invention, the deviation is chosen from an absolute deviation and a relative deviation.

[0026] In a third embodiment of the first aspect of the invention, the machine learning model is determined by a supervised machine learning technique.

[0027] In a fourth embodiment of the first aspect of the invention, the data processing system further comprises at least one memory which is coupled to the processor.

[0028] In the fourth embodiment of the first aspect of the invention, the processor is further designed to save the vector in memory, before providing it as input to the trained machine learning model.

[0029] Finally, in the fourth embodiment of the first aspect of the invention, the processor is further designed to determine whether it is appropriate to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

[0030] In a fifth embodiment of the first aspect of the invention, the data processing system further comprises an interactive manipulation surface.

[0031] A second aspect of the invention relates to a computer-implemented method for evaluating the fine motor skills of an individual in the context of carrying out a routine activity which mobilizes the graphomotor gesture in the execution of at least one drawing on an interactive manipulation surface.

[0032] In particular, the evaluation is carried out on the basis of a single current evaluation model which is chosen from a plurality of evaluation models.

[0033] In practice, each evaluation model describes, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators which are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index which is representative of a combination of at least two kinematic and / or kinetic indicators.

[0034] The method includes a first step of providing at least one sensor which is coupled to the interactive manipulation surface.

[0035] The method then comprises a second step of supplying at least one processor

[0036] The process then comprises a detection and recording step, by the sensor, of at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when performing at least one tracing within the activity, the sequence of movements comprising a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes

[0037] The method then comprises a first step of calculation, by the processor, of a plurality of kinematic and / or kinetic indicators from the attributes which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements

[0038] The method then comprises a first step of execution, by the processor, of the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index

[0039] The method then includes a second step of calculation, by the processor, of at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model

[0040] The process then includes a step of forming, by the processor, a vector which includes a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index, and the calculated gap.

[0041] The method then includes a third step of providing the input vector to a machine learning model trained to predict the next activity evaluation model.

[0042] The process finally includes a second step of execution, by the processor, of the machine learning model trained so as to obtain, at the output of the machine learning model, a third piece of information which is representative of the next evaluation model to be used to evaluate the activity.

[0043] In one embodiment of the second aspect of the invention, the method further comprises a fourth step of supplying a memory which is coupled to the processor.

[0044] Then, in the embodiment of the second aspect of the invention, the method then includes a step of saving, by the processor, the vector in memory, before providing it as input to the trained machine learning model.

[0045] Finally, in the embodiment of the second aspect of the invention, the method also includes a step of determining, by the processor, whether it is appropriate to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

[0046] A third aspect of the invention relates to a computer-implemented method for the supervised training of a machine learning model intended to predict the next evaluation model of a routine activity that mobilizes the graphomotor gesture in the execution of at least one drawing on an interactive manipulation surface.

[0047] In particular, the evaluation is carried out on the basis of a single current evaluation model which is chosen from a plurality of evaluation models.

[0048] In practice, each evaluation model describes, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators which are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index which is representative of a combination of at least two kinematic and / or kinetic indicators.

[0049] The process includes a first step of supplying at least one processor.

[0050] The method then includes a step of obtaining, by the processor, a plurality of movement sequence history segments, each being produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when carrying out at least one tracing within the context of the activity, the plurality of movement sequence history segments comprising a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes.

[0051] The method then includes a first step of calculation, by the processor, of a plurality of kinematic and / or kinetic indicators from the attributes which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements.

[0052] The method then includes a step of execution, by the processor, of a first current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index.

[0053] The method then includes a second step of calculation, by the processor, of at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model.

[0054] The process then includes a first step of formation, by the processor, of a first vector which includes a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index, and the calculated gap.

[0055] The process then comprises a second step of training, by the processor, a second vector which includes a fourth piece of information which is representative of a second evaluation model, the second evaluation model being obtained by evaluation by at least one expert to determine whether the first activity evaluation model should be replaced by the second evaluation model.

[0056] The method finally includes a step of supervised training, by the processor, of a learning model that includes at least one input designed to receive the first vector, and at least one output designed to receive the second vector. Brief description of the drawings

[0057] Other features and advantages of the invention will be better understood from the following description and with reference to the accompanying drawings, given by way of illustration and not limitation.

[0058] [Fig-1] The [Fig.1] represents a system according to the invention.

[0059] [Fig.2] Fig.2 represents a first method according to the invention.

[0060] [Fig.3] Fig.3 represents a second method according to the invention.

[0061] The figures do not necessarily respect the scales, particularly in thickness, and This is for illustrative purposes only. Description of the implementation methods

[0062] Summary presentation of the invention

[0063] One of the objectives of this invention is to provide a solution that allows an objective and precise evaluation of the graphomotor gesture.

[0064] To this end, the inventors propose to evaluate, separately, the different facets of the graphomotor gesture produced in the context of an activity which is carried out on an interactive manipulation surface.

[0065] In practice, each facet of the graphomotor gesture is evaluated by a respective evaluation model.

[0066] Then, for a current activity which is evaluated by the current evaluation model, a machine learning model which has been trained to predict the next evaluation model to be used, based on the results obtained by the current evaluation model.

[0067] Advantage of the invention

[0068] In this way, professionals involved in education and health will benefit from a reliable method for assessing fine motor skills without being influenced by their own subjectivity.

[0069] The evaluation system

[0070] As illustrated in [Fig.1], the invention relates to a data processing system 100 for evaluating the fine motor skills of an individual.

[0071] As is well known, "fine motor skills" refers to the set of precise and controlled movements of small parts of the human body, such as the fingers, hands, and wrists. These movements are important in many daily activities. Daily activities that require fine motor skills, such as writing, drawing, sewing, or handling delicate objects, are essential for developing fine motor skills and performing tasks with precision and efficiency.

[0072] In practice, in the invention, the individual carries out a routine activity which mobilizes the graphomotor gesture in such a way as to produce at least one trace, possibly residual, on an interactive manipulation surface 110.

[0073] The term "activity which mobilizes graphomotor gesture" means any action or task which requires the intentional use of fine motor skills of the hands, fingers and wrist to make at least one mark on an interactive manipulation surface.

[0074] For example, these activities may include grasping exercises, construction games, thinking games, coloring, handwriting and all other activities that require eye-hand coordination.

[0075] However, depending on the needs and resources available, other activities may be considered, without requiring substantial modifications to the invention.

[0076] An “interactive manipulation surface” is understood to mean an interface (e.g. in 2D or 3D) that allows an individual to interact with graphic or digital elements using a variety of interactive tools such as fingers, styluses and other input devices.

[0077] For example, an interactive manipulation surface can be chosen from among the touch screens of smartphones and tablets, interactive whiteboards, gesture interaction surfaces and motion tracking devices.

[0078] However, depending on the needs and resources available, other interactive manipulation surfaces may be considered, without requiring substantial modifications to the invention.

[0079] In practice, in the invention, the activity is evaluated on the basis of a single current evaluation model which is chosen from a plurality of evaluation models.

[0080] The term "evaluation model" means a set of rules or criteria that allow us to measure and judge the quality, effectiveness or performance of at least one routine activity that involves graphomotor gestures.

[0081] In practice, in the invention, each evaluation model describes, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators which are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index which is representative of a combination of at least two kinematic and / or kinetic indicators.

[0082] Thus, the invention may include an evaluation model that describes, for two or more than two given activities, a relationship between, on the one hand, two or more kinematic and / or kinetic indicators and, on the other hand, two or more composite indices.

[0083] A "kinematic indicator" is understood to mean a measurable physical quantity which describes the movement of a graphomotor gesture without taking into account the forces which cause it such as the resistance of the interactive manipulation surface 110 and the muscular tension of the individual when carrying out the tracing.

[0084] For example, the kinematic indicator can be chosen from the length of the path, the number of pauses and the total execution time.

[0085] A "kinetic indicator" is understood to mean a measurable physical quantity which describes the movement of a graphomotor gesture while taking into account the forces which cause it such as the resistance of the interactive manipulation surface 110 and the muscular tension of the individual when carrying out the tracing.

[0086] For example, the kinetic indicator can be chosen from among the pressure exerted on the interactive manipulation surface 110, the direction of the trace, the inclination of the trace, the speed of tracing and the acceleration(s) / deceleration(s).

[0087] In the invention, depending on the needs and available resources, other indicators may be considered without requiring substantial modifications to the invention. For example, these indicators may provide information on the result of a graphomotor gesture, the characteristics of the gesture that gave rise to it, the underlying cognitive programming and control processes at work during the execution of this gesture, as well as on the profile of the person performing it.

[0088] In a first particular implementation, the composite index is calculated according to a predetermined mathematical formula which weights and combines kinematic and / or kinetic indicators.

[0089] In a second particular implementation, the composite index is calculated according to a predetermined algorithm which manipulates kinematic and / or kinetic indicators.

[0090] In the invention, the data processing system 100 comprises at least one sensor 120 and at least one processor 130.

[0091] Thus, the system 100 can include two or more of two sensors 120 and two or more of two processors 130.

[0092] In a particular implementation of the system 100, it further comprises at least one interactive manipulation surface 110 as described above.

[0093] Thus, in this particular implementation, the system 100 can comprise two or more of two interactive manipulation surfaces 110.

[0094] In the invention, the sensor 120 is coupled to the interactive manipulation surface 110.

[0095] In addition, the sensor 120 is designed to detect and record at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory (e.g. a stylus, a mouse) that is usable by the individual when performing at least one tracing within the context of the activity.

[0096] Thus, the sensor 120 can detect and record two or more sequences of movements produced by two or more fingers of the individual or by two or more pointing accessories that are usable by the individual when making at least one trace in the context of the activity.

[0097] In one example of the sensor 120, it comprises a combination of sensors which are selected from, a capacitive screen, a resistive screen, an accelerometer, a gyroscope, a pressure sensor, an infrared sensor and an ultrasonic sensor.

[0098] However, depending on the needs and resources available, other sensors may be considered, without requiring substantial modifications to the invention.

[0099] In the invention, the sequence of movements comprises a plurality of sequential trajectory points.

[0100] In practice, each trajectory point of the sequence of movements is associated with at least a plurality of attributes.

[0101] For example, the attributes of a point may include spatial data (e.g. its x, y and possibly z coordinates) relative to the interactive manipulation surface 110, temporal data relative to a predetermined temporal reference frame, and kinesthetic data (e.g. position, orientation, velocity, acceleration and forces exerted during these movements).

[0102] However, depending on the needs and resources available, other attributes may be considered, without requiring substantial modifications to the invention.

[0103] In the invention, the processor 130 is coupled to the sensor 120.

[0104] Furthermore, the processor 130 is designed to calculate a plurality of kinematic and / or kinetic indicators from the attributes that are associated with all or part of the plurality of sequential trajectory points of the sequence of movements.

[0105] In addition, the processor 130 is also designed to execute the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index.

[0106] Also in the invention, the processor 130 is also designed to calculate at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model.

[0107] Thus, the processor 130 can calculate two or more of two deviations between the composite index and two or more of two predetermined thresholds which are associated with the current evaluation model.

[0108] In a particular first implementation, the predetermined threshold includes a success threshold and / or failure threshold.

[0109] In other words, the predetermined threshold may include either only a success threshold, or only a failure threshold, or both a success threshold and a failure threshold.

[0110] The term "success threshold" means the level that the individual must reach to be considered as having succeeded in the activity.

[0111] The term "failure threshold" means the level below which the individual cannot be considered to have succeeded in the activity.

[0112] In a second particular implementation, the deviation is chosen from an absolute deviation and a relative deviation.

[0113] The term “absolute deviation” means the difference between the composite index and the predetermined threshold.

[0114] The term "relative deviation" means the difference between the composite index and the predetermined threshold expressed as a percentage of one of these two values.

[0115] In the invention, the processor 130 is also designed to form a vector which includes a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index and the calculated gap.

[0116] In a known manner, a "vector" is understood to be a data structure that stores contiguous elements in memory.

[0117] In addition, the processor 130 is also designed to provide the input vector for a machine learning model trained to predict the next activity evaluation model.

[0118] A machine learning model is a type of computer program that learns to perform a task by training on data rather than being explicitly programmed for that task. In practice, a machine learning model uses algorithms that identify patterns in data and use them to make predictions or decisions about new data.

[0119] In a particular implementation, the machine learning model is determined by a supervised machine learning technique.

[0120] In one example, the supervised machine learning technique is chosen from linear regression, regression decision tree, regression support vector machine (SVM) and feed-forward neural network with a final layer that presents a single output unit without nonlinear activation.

[0121] However, depending on the needs and available resources, other supervised machine learning techniques may be considered, without requiring substantial modifications of the invention.

[0122] In the invention, the processor 130 is also designed to run the trained machine learning model so as to obtain, as output from the trained machine learning model, a third piece of information which is representative of the next evaluation model to be used to evaluate the activity.

[0123] In one embodiment of the invention, the system 100 further comprises at least one memory 140 which is coupled to the processor 130.

[0124] Thus, the system 100 can include two or more of two memories 140.

[0125] In this embodiment, the processor 130 is further designed to save the vector in memory 140, before providing it as input to the trained machine learning model.

[0126] Then, in this embodiment, the processor 130 is further designed to determine whether it is appropriate to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

[0127] The evaluation process

[0128] As illustrated in [Fig.2], the invention also relates to a computer-implemented method 200 for evaluating an individual's fine motor skills in the context indicated above.

[0129] In the invention, the method 200 includes a first step of supplying 210 with at least one sensor 120 which is coupled to the interactive manipulation surface 110.

[0130] Thus, the invention may include the provision of two or more of two sensors 120.

[0131] In the invention, the method 200 includes a second step of supplying 220 with at least one processor 130.

[0132] Thus, the invention may include the provision of two or more of two processors 130.

[0133] In the invention, the method 200 includes a detection and recording step 230, by the sensor 120, of at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when carrying out at least one tracing in the context of the activity.

[0134] Thus, the invention may include the detection and recording of two or more sequences of movements produced by two or more fingers of the individual or by two or more pointing accessories that are usable by the individual when making at least one drawing in the context of the activity.

[0135] As indicated above, the sequence of movements comprises a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes.

[0136] In the invention, the method 200 comprises a first calculation step 240, by the processor 130, of a plurality of kinematic and / or kinetic indicators from the attributes which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements.

[0137] In the invention, the method 200 includes a first execution step 250, by the processor 130, of the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index.

[0138] In the invention, the method 200 includes a second calculation step 260, by the processor 130, of at least one deviation between the composite index and at least one predetermined threshold which is associated with the current evaluation model.

[0139] Thus, the invention may include the calculation of two or more deviations between the composite index and two or more predetermined thresholds that are associated with the current evaluation model.

[0140] In the invention, the process 200 includes a training step 270, by the processor 130, of a vector which includes a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index and the calculated gap.

[0141] In the invention, the method 200 includes a third step 280 of supplying the input vector to a machine learning model trained to predict the next activity evaluation model.

[0142] In the invention, the method 200 includes a second execution step 290, by the processor 130, of the machine learning model trained so as to obtain, at the output of the machine learning model, a third piece of information which is representative of the next evaluation model to be used to evaluate the activity.

[0143] In one embodiment of the process 200, this includes a fourth step of supplying 291 a memory 140 which is coupled to the processor 130.

[0144] Then, in this embodiment of the process 200, it includes a step of saving 292, by the processor 130, the vector in the memory 140, before providing it as input to the trained machine learning model.

[0145] Finally, in this embodiment of the process 200, it includes a determination step 293, by the processor 130, of whether it is appropriate to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

[0146] The training method

[0147] As illustrated in [Fig. 3], the invention also relates to a 300 im method computer-implemented for supervised training of a machine learning model intended to predict the next evaluation model of a routine activity that involves graphomotor gestures in the context indicated above.

[0148] In the invention, the process 300 includes a first step of supplying 310 with at least one processor 130.

[0149] Thus, the invention may include the provision of two or more of two processors 130.

[0150] In the invention, the method 300 includes a step of obtaining 320, by the processor 130, a plurality of movement sequence history segments.

[0151] In practice, each movement sequence is produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when making at least one mark in the context of the activity.

[0152] In particular, the plurality of movement sequence history segments comprises a plurality of trajectory sequential points, each point of which is associated with at least a plurality of attributes.

[0153] Thus, the invention may include obtaining a plurality of movement sequence history segments, each being produced by two or more of the individual's fingers or by two or more of pointing accessories that are usable by the individual when performing at least one tracing within the activity.

[0154] In the invention, the method 300 comprises a first calculation step 330, by the processor 130, of a plurality of kinematic and / or kinetic indicators from the attributes which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements.

[0155] In the invention, the method 300 includes an execution step 340, by the processor 130, of a first current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index.

[0156] In the invention, the method 300 includes a second calculation step 350, by the processor 130, of at least one deviation between the composite index and at least one predetermined threshold which is associated with the current evaluation model.

[0157] Thus, the invention may include the calculation of two or more deviations between the composite index and two or more predetermined thresholds that are associated with the current evaluation model.

[0158] In the invention, the process 300 includes a first training step 360, by the processor 130, of a first vector which includes a first piece of information which is representative of the activity, a second piece of information which is representative of the current evaluation model, the composite index and the calculated gap.

[0159] In the invention, the method 300 includes a second training step 370, by the processor 130, of a second vector which includes a fourth piece of information which is representative of a second evaluation model, the second evaluation model being obtained by evaluation by at least one expert to determine whether it is appropriate to replace the first evaluation model of the activity with the second evaluation model.

[0160] Thus, the invention may include the evaluation by two or more experts to determine whether it is appropriate to replace the first activity evaluation model with the second evaluation model.

[0161] In the invention, the method 300 includes a supervised training step 380, by the processor 130, of a learning model which includes at least one input designed to receive the first vector and at least one output designed to receive the second vector.

[0162] We have described and illustrated the invention. However, the invention is not limited to the embodiments we have presented. Thus, an expert in the field may deduce other variations and embodiments from the description and accompanying figures.

[0163] The invention can be the subject of numerous variations and applications other than those described above. In particular, unless otherwise indicated, the various structural and functional features of each of the embodiments described above should not be considered as combined and / or closely and / or inextricably linked to one another, but rather as mere juxtapositions. Furthermore, the structural and / or functional features of the various embodiments described above may be the subject, in whole or in part, of any different juxtaposition or any different combination.

[0164] When the description indicates that an element is "designed" for a given function, this means that the element is created specifically for the purpose of fulfilling that desired function.

Claims

Demands

1. A data processing system (100) for evaluating an individual's fine motor skills in the context of performing a routine activity that involves graphomotor gestures in creating at least one drawing on an interactive manipulation surface (110), the evaluation being carried out on the basis of a single routine evaluation model chosen from a plurality of evaluation models, each describing, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators that are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index that is representative of a combination of at least two kinematic and / or kinetic indicators, the data processing system (100) comprising, - at least one sensor (120) that is coupled to the interactive manipulation surface (110),the sensor (120) being designed to detect and record at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when performing at least one tracing within the context of the activity, the sequence of movements comprising a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes, - at least one processor (130) which is coupled to the sensor (120) and which is designed to, - calculate a plurality of kinematic and / or kinetic indicators from the attributes which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements, - execute the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index,- calculate at least one difference between the composite index and at least one predetermined threshold associated with the current evaluation model, - form a vector that includes: - a first piece of information representative of the activity, - a second piece of information representative of the current evaluation model, - the composite index, and - the calculated gap, - provide the input vector to a machine learning model trained to predict the next activity evaluation model, - run the trained machine learning model in such a way as to obtain, as output of the trained machine learning model, a third piece of information that is representative of the next evaluation model to be used to evaluate the activity.

2. Data processing system (100) according to claim 1, wherein the predetermined threshold includes a success threshold and / or a failure threshold.

3. Data processing system (100) according to any one of claims 1 to 2, wherein the deviation is chosen from an absolute deviation and a relative deviation.

4. Data processing system (100) according to any one of claims 1 to 3, wherein the machine learning model is determined by a supervised machine learning technique.

5. A data processing system (100) according to any one of claims 1 to 4, further comprising at least one memory (140) which is coupled to the processor (130), and wherein the processor (130) is further designed to: - save the vector in the memory (140), before providing it as input to the trained machine learning model, and - determine whether to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

6. Data processing system (100) according to any one of claims 1 to 5, further comprising at least one interactive manipulation surface (110).

7. A computer-implemented method (200) for assessing an individual's fine motor skills in the context of performing a routine activity that involves graphomotor gestures in creating at least one drawing on an interactive manipulation surface (110), the assessment being carried out on the basis of a single routine assessment model chosen from a plurality of assessment models, each describing, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators that are respectively representative of all or part of at least one a sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index that is representative of a combination of at least two kinematic and / or kinetic indicators, the method (200) comprising, - a first step of supplying (210) at least one sensor (120) which is coupled to the interactive manipulation surface (110), - a second supply step (220) of at least one processor (130), - a detection and recording step (230), by the sensor (120), of at least one sequence of movements produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when carrying out at least one tracing within the context of the activity, the sequence of movements comprising a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes, - a first calculation step (240), by the processor (130), of a plurality of kinematic and / or kinetic indicators from the attributes that are associated with all or part of the plurality of sequential trajectory points of the sequence of movements, - a first execution step (250), by the processor (130), of the current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index, - a second calculation step (260), by the processor (130), of at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model, - a training step (270), by the processor (130), of a vector which includes, - initial information that is representative of the activity, - a second piece of information that is representative of the current evaluation model - the composite index, and - the calculated difference, - a third step of supplying (280) the input vector to a machine learning model trained to predict the next activity evaluation model, - a second execution step (290), by the processor (130), of the machine learning model trained in such a way as to obtain, in output of the trained machine learning model, a third piece of information that is representative of the next evaluation model to be used to assess the activity.

8. A method (200) according to claim 7, further comprising: - a fourth step of supplying (291) a memory (140) which is coupled to the processor (130), - a step of saving (292), by the processor (130), the vector in the memory (140), before supplying it as input to the trained machine learning model, and - a step of determining (293), by the processor (130), whether it is appropriate to propose a new activity to the individual, based on pairs which are each made up of a composite index and the associated gap, the pairs being extracted from at least a predetermined number of vectors.

9. A computer-implemented method (300) for the supervised training of a machine learning model intended to predict the next evaluation model of a current activity that mobilizes graphomotor gestures in the execution of at least one drawing on an interactive manipulation surface, the activity being evaluated by a current evaluation model chosen from a plurality of evaluation models, each evaluation model describing, for at least one given activity, a relationship between, on the one hand, at least two kinematic and / or kinetic indicators that are respectively representative of all or part of at least one sequence of movements associated with a graphomotor gesture and, on the other hand, at least one composite index that is representative of a combination of at least two kinematic and / or kinetic indicators, the method (300) comprising, - a first step of supplying (310) at least one processor (130),- a step of obtaining (320), by the processor (130), a plurality of movement sequence history segments, each produced by at least one finger of the individual or at least one pointing accessory that is usable by the individual when performing at least one tracing within the context of the activity, the plurality of movement sequence history segments comprising a plurality of sequential trajectory points, each point of which is associated with at least a plurality of attributes, - a first calculation step (330), by the processor (130), of a plurality of kinematic and / or kinetic indicators from the attributes, which are associated with all or part of the plurality of sequential trajectory points of the sequence of movements, - an execution step (340), by the processor (130), of a first current evaluation model from all or part of the plurality of kinematic and / or kinetic indicators, so as to obtain at least one composite index, - a second calculation step (350), by the processor (130), of at least one difference between the composite index and at least one predetermined threshold which is associated with the current evaluation model, - a first training step (360), by the processor (130), of a first vector which includes, - initial information that is representative of the activity, - a second piece of information that is representative of the current evaluation model - the composite index, and - the calculated difference, - a second training step (370), by the processor (130), of a second vector which includes a fourth piece of information which is representative of a second evaluation model, the second evaluation model being obtained by evaluation by at least one expert to determine whether it is appropriate to replace the first evaluation model of the activity with the second evaluation model, - a supervised training step (380), by the processor (130), of a learning model which includes, - at least one input designed to receive the first vector, and - at least one output designed to receive the second vector.

Images