Method, device and computer program for evaluating a movement sequence

Abstract

The invention relates to a method for evaluating a movement sequence performed by a patient influenced by an orthopaedic device, wherein the method comprises the following steps: - providing digital video data, which contain a recording of the patient influenced by the orthopaedic device while performing the specified movement sequence, to an electronic computing unit; - determining, by means of the electronic computing unit, a plurality of patient-related movement features of the movement sequence of the patient contained in the digital video data; - assigning, by means of the electronic computing unit, the movement sequence contained in the digital video data to a movement-sequence type depending on the determined patient-related movement features; - establishing, on the basis of the determined patient-based movement features, and by means of the electronic computing unit, whether the movement sequence contained in the digital video data corresponds to a predefined degree of correctness, in view of the associated movement-sequence type based on its execution; and - determining, by means of the computing unit, at least one evaluation characteristic on the basis of the patient-based movement features.

Description

[0001] Method, device and computer program for evaluating a movement sequence

[0002] The invention relates to a method for evaluating a movement sequence performed by a patient fitted with an orthopedic device. The invention also relates to a device and a computer program for this purpose.

[0003] For the purposes of the present invention, orthopedic devices (also referred to as fittings) include, in particular, orthoses, prostheses, exoskeletons, and possibly also wheelchairs. Orthoses are products that support, assist, protect, and / or restrict the freedom of movement of a patient's body part, such as a joint, in order to prevent excessive strain. Prostheses, in particular exoskeletons, on the other hand, do not replace, or no longer replace, the patient's body parts. Exoskeletons are, in particular, mechanical support structures designed to support, assist, and / or protect the patient's main musculoskeletal system.

[0004] An orthopedic device also includes muscle-stimulating suits or coverings in which various muscles are stimulated externally using electrodes. Such a muscle-stimulating suit helps patients with spasticity by stimulating the antagonistic muscles and thus relaxing the spastic muscle (the antagonist of the stimulated muscle) via the physiological reflex mechanism. The effect produced by the muscle-stimulating suit and its application lasts for a certain time after application, so that with regular use, improvements can be seen even without the orthopedic device. The orthopedic device therefore influences the patient. Such a suit can be a full-body suit or just a partial suit. Other stimulating orthoses are also included in the term, such as a cuff for stimulating individual muscles or nerve pathways.can be used for muscle weakness. This can be used, for example, to treat foot drop caused by muscle weakness.

[0005] In the following, a patient is defined as any user of the orthopedic device. This refers to the wearer of the orthopedic device.

[0006] Every orthotic device is attached to a part of the patient's body. It does not necessarily have to come into contact with the patient's skin. Orthoses and exoskeletons, for example, are often worn over clothing, so that the clothing, such as trousers, is between the orthotic device and the patient's skin. However, a knee orthosis, for example, is attached to the patient's knee or leg. A prosthesis always has an interface element connected to an amputation stump or another body part, which is attached to the respective body part. A leg prosthesis, for example, uses a prosthetic socket that represents the interface between the prosthesis and the amputation stump. In this case, the amputation stump would be the patient's body part.A liner is typically used between the skin surface of the patient's amputation stump and the prosthetic socket to reduce shear forces acting on the skin.

[0007] Once the patient has worn the orthotic device for a while, it is often desirable to evaluate the outcome or effect of the device on the patient. This allows for timely counteraction to any misuse or insufficiently precise configuration of the device.

[0008] For this purpose, so-called Outcome Measurement Tests (OMTs) are often performed. These tests require the patient to perform a specific movement or sequence of movements while wearing the orthotic device. Specific movement characteristics of the movement sequence performed by the patient are then determined, which then provides an assessment of how well or poorly the patient performed the test. This assessment then determines whether wearing the orthotic device is achieving the desired effect, whether a different device should be selected, or whether the configuration of the existing device needs to be changed, varied, and / or adapted.

[0009] A well-known outcome measurement test is the "Timed Up and Go Test." The patient sits on a chair and, upon a command, must stand up, walk a certain distance (usually 3 meters) at a normal and safe pace, then turn around, walk back to the chair, and sit down again. The time it takes the patient to go from standing up to sitting down is measured. If such an OMT is repeated at regular intervals, the time measurement can be used to evaluate the effectiveness of the orthopedic device.

[0010] Another OMT is the "6-minute walk test," in which the patient walks for 6 minutes, with the result being the distance covered. The patient can take as many breaks as they wish. The evaluation of the orthopedic device is then based on the distance covered in 6 minutes.

[0011] In practice, however, such OMTs are rarely performed with the necessary precision. The reason for this is that performing such tests represents a significant additional burden on the daily care process for care providers, such as medical supply stores or orthotists. These OMTs compete with other care-related processes, so they are often omitted. As a result, the effectiveness of the orthotist device is not evaluated at regular intervals, potentially prolonging recovery and delaying or even failing to implement necessary adjustments to the orthotist device.This not only leads to additional stress for the patient, but also to higher treatment costs, as the patient has to be cared for longer than is actually necessary or receives suboptimal orthopaedic equipment that has to be replaced again later.

[0012] Due to the significant contribution that OMTs make to patient recovery and improved well-being, there is a need in everyday orthopedic and therapeutic practice to perform OMTs comprehensively and at short intervals. However, due to the current practice of reliably performing such OMTs within healthcare facilities, this cannot be achieved with current practice.

[0013] EP 3 786 971 A1 discloses a training device that records a training exercise using a mobile device and evaluates the performance of the exercise. The user can be informed of how precisely the training exercise is performed.

[0014] From US 2021 / 0225195 A1, a method for improving gymnastics exercises is known in which a movement sequence is also recorded and evaluated by a camera.

[0015] From EP 3 191 989 B1 a method for identifying a motor task of a user is known, wherein here too a video of the motor movement is recorded and analyzed in order to classify the recorded movement.

[0016] It is therefore an object of the present invention to provide an improved method with which the evaluation of an orthopaedic device with regard to its effect on the patient can be improved.

[0017] This object is achieved according to the invention with the method for evaluating a movement sequence performed by a patient influenced by an orthopedic device, as claimed in claim 1. Advantageous embodiments of the invention can then be found in the corresponding subclaims. According to claim 1, a method for evaluating a movement sequence is proposed, wherein the movement sequence is performed by a patient influenced by an orthopedic device. Such a movement sequence is, in particular, an OMT, with which the effectiveness of the orthopedic device with which the patient is provided or by which the patient is influenced is to be evaluated.

[0018] The procedure includes the following steps

[0019] - Providing digital video data containing a recording of the patient influenced by the orthopaedic device during the execution of the specified movement sequence to an electronic processing unit,

[0020] - Determining several patient-related movement characteristics of the patient's movement sequence contained in the digital video data by means of the electronic processing unit,

[0021] - Assigning the movement sequence contained in the digital video data to a movement sequence type depending on the determined patient-related movement characteristics by means of the electronic processing unit,

[0022] - Determining, by means of the electronic processing unit, whether the movement sequence contained in the digital video data corresponds to a predefined degree of correctness with regard to the assigned movement sequence type in relation to its execution, whereby the determination is made depending on the determined patient-related movement characteristics, and

[0023] - Determining at least one evaluation key figure depending on the patient-related movement characteristics by means of the computing unit.

[0024] According to the invention, an automated evaluation of such a movement sequence, in particular an OMT, is proposed. First, digital video data is provided containing a recording of the patient performing the specified movement sequence. Such a video can be recorded, for example, using a mobile device such as a smartphone, tablet, or computer camera. The recording can be made by the patient themselves, without having to visit an orthopedic technician or a medical supply store. Rather, the video, which contains the specified movement sequence as a sequence of individual images, can be recorded at home in a familiar environment.The video recorded in this way with the digital video data is now transmitted from the video recording device to an electronic processing unit, so that the digital video data with the recorded movement sequence is made available to this electronic processing unit.

[0025] The video data shows the patient performing the specified movement sequence. The patient is influenced by their orthopedic device while executing the movement sequence. This can mean that the patient wears the orthopedic device or is equipped with the orthopedic device while performing the specified movement sequence. It is also conceivable, however, that the patient is previously influenced by the orthopedic device, for example, by a muscle-stimulating suit, and then performs the movement sequence. In this case, the patient is influenced by the muscle-stimulating suit that previously acted on their body, so that the effect of such a suit on the patient can be examined and evaluated.

[0026] The influence of an orthopaedic device is therefore understood not only to mean the provision of the orthopaedic device, but also the effect of an orthopaedic device on the patient through external stimuli, both during and after use.

[0027] The video recording device and the electronic processing unit can be integrated into a single device. If the video recording device is a smartphone, for example, all process steps can be performed on the smartphone. It is also possible, however, for the video recording device to be connected to a data network via an interface, for example, via LAN or WLAN, with the video recording device then transmitting the recorded video data to the electronic processing unit, which is also connected to the network.

[0028] The patient has specific instructions regarding how to perform the specified movement sequence. The electronic processing unit, however, usually does not know which movement sequence or movement type the patient has performed. Thus, the electronic processing unit generally does not know which OMT the patient is performing in the video data.

[0029] First, several patient-related movement characteristics of the movement sequence contained in the digital video data are determined using the electronic processing unit. Based on these patient-related movement characteristics, the movement sequence performed by the patient in the video is then assigned to a movement sequence type, for example, a specific OMT. The electronic processing unit thus automatically determines the underlying movement sequence type performed by the patient in the video.

[0030] Such patient-related movement characteristics can, for example, be the joint angle of a patient's joint at a specific point in time or as a progression over time. For example, the joint angle of a joint as a progression over time can be used to determine whether the patient is currently rising from a chair or sitting down, allowing a specific movement pattern to be inferred.

[0031] Once the movement sequence type has been determined, the next step is for the electronic processing unit to determine whether the movement sequence contained in the digital image data meets a predefined accuracy level in terms of its execution. This ensures that the movement sequence has been executed correctly enough to make an evaluation possible and meaningful. The determination of whether the movement sequence meets the predefined accuracy level is based on the patient-related movement characteristics previously extracted from the video data.

[0032] If it has been determined that the movement sequence corresponds to a predefined correctness level, an evaluation score is then determined by the computing unit based on the patient-related movement characteristics. This evaluation score indicates how qualitatively the movement sequence was performed by the patient using the orthopedic device. The evaluation score can also incorporate characteristics or properties relating to the movement sequence itself, such as the distance covered or the time the patient needs to cover a certain distance. The evaluation score is therefore preferably determined not only based on the patient-related movement characteristics, but also on execution-related movement information, such as time or distance covered.

[0033] With the help of the present invention, it is thus possible to reduce the time required to carry out an OTM due to a high degree of automation so that these OTMs can be carried out in the everyday care process.

[0034] It is particularly advantageous if the configuration of the orthopedic device is adjusted based on the determined evaluation index. This can involve selecting another suitable orthopedic device or adapting the existing device. For example, the structure of the device can be optimized for the patient, or the device's setting parameters, particularly in microprocessor-controlled devices, such as the damping curve of a knee joint, can be adjusted. A suggestion for adjusting the orthopedic device can also be provided automatically based on the evaluation index and / or the determined movement characteristics.

[0035] According to one embodiment, a first machine learning system is provided on the electronic computing unit, which first machine learning system has learned a relationship between movement sequences contained in digital video data of patients influenced by an orthopedic device as input data and physical characteristics of the patient as output data, wherein in order to determine physical characteristics of the patient in the movement sequence, the digital video data are fed as input data to the first machine learning system and the physical characteristics of the patient are then obtained as output data, wherein the patient-related movement characteristics are identified as a function of the determined physical characteristics of the patient.For example, the machine learning system can be used to identify joint points of a joint and / or body parts of the patient in the movement sequence of digital video data, whereby based on these joint points and / or body parts, the patient-related movement characteristics, such as joint angle and / or change in the joint angle over time, can then be determined.

[0036] According to one embodiment, the machine learning system is an artificial neural network.

[0037] The machine learning system can be trained by presenting digital video data and / or individual images as training data to the machine learning system, in which the relevant body characteristics have already been identified, i.e., for example, the joint points and / or body parts are already known from the video data and / or individual images. A large number of such training data are presented to the machine learning system, enabling corresponding machine learning. It has proven advantageous to use training data that also contains orthopedic devices, since movement sequences and body contours can sometimes differ considerably due to the use of orthopedic devices.

[0038] According to one embodiment, it is provided that at least one joint point of a joint of the patient in the movement sequence, a joint angle of at least one joint of the patient in the movement sequence, a relative position between at least two joints of the patient and / or a silhouette of the patient's body as well as changes in these body characteristics over time are identified as a body characteristic.

[0039] It is conceivable that the body characteristics that can be determined from the first machine learning system are at least partially identical to the patient-related movement characteristics, although this does not necessarily have to be the case. According to one embodiment, it is provided that the patient-related movement characteristics

[0040] - a joint angle of a patient's joint at a specific point in time during the movement sequence or as a progression over time,

[0041] - a relative position between at least two joints of the patient at a specific point in time during the movement sequence or as a progression over time, and / or

[0042] - a touchdown and / or a setdown point is determined in the movement sequence or as a progression over time.

[0043] According to one embodiment, a second machine learning system is provided on the electronic computing unit, which has learned a relationship between patient-related movement characteristics of the patient when performing a predetermined movement sequence as input data and a movement sequence type of the respective movement sequence as output data, wherein for the assignment of the movement sequence contained in the digital video data to a corresponding movement sequence type, the determined patient-related movement characteristics are fed to the second machine learning system and the movement sequence type is then obtained as output data.

[0044] Thus, based on the patient-related movement characteristics determined, for example, from the first machine learning system, the movement sequence type can be determined by inputting the patient-related movement characteristics, such as joint angles in absolute terms or as a progression over time, into the second machine learning system and outputting the movement sequence type associated with this as output data.

[0045] Such a machine learning system can also be an artificial neural network.

[0046] The second machine learning system can also be presented with training data to learn the relationship, which contains patient-related movement characteristics and the associated movement sequence type. This means that the training data knows the association of patient-related movement characteristics with the underlying movement sequence type and can thus, for example, train the second machine learning system via appropriate feedback learning.

[0047] According to one embodiment, an assignment rule set is provided on the electronic computing unit, which contains an assignment of certain combinations of patient-related movement characteristics of the patient when performing a predetermined movement sequence as input data to a respective movement sequence type as output data, wherein for the assignment of the movement sequence contained in the digital video data to a corresponding movement sequence type, the determined patient-related movement characteristics are fed to the assignment rule set and the movement sequence type is then obtained as output data.

[0048] Such an assignment rule set recognizes specific combinations of patient-related movement characteristics and assigns them to a specific movement sequence type in which this combination of patient-related movement characteristics always occurs. Such an assignment rule set has the advantage that no training data needs to be provided for training a machine learning system, as required. It can also be provided that not only the movement sequence type is determined, but also a probability is output that indicates how likely the assignment to the movement sequence type is.

[0049] According to one embodiment, it is provided that the correctness measure corresponds to or is derived from a probability in the assignment to the movement sequence type.

[0050] This means that the probability of assignment to the movement sequence type, for example, using a machine learning system or assignment rule set, is used to check whether the execution of the movement sequence meets the correctness standard. The more probable the assignment, the more correctly the movement was executed. The more uncertain the assignment, the more correctly the movement was executed. According to one embodiment, the evaluation index is determined by the computing unit as a function of a probability of assignment to the movement sequence type.

[0051] In this embodiment, the probability of assignment to the movement sequence type is included in the determination of the evaluation key figure, so that the evaluation key figure is higher, the more probable the assignment to the movement sequence type is.

[0052] According to one embodiment, it is provided that the evaluation index is determined as a function of execution-related movement information of the movement sequence derived from the patient-related movement characteristics.

[0053] Such execution-related movement information can be, for example, the distance traveled or the time required to cover a given distance.

[0054] According to one embodiment, the movement information includes, among other things, the duration of the movement, the presence of a specific movement pattern, information about reaching specific joint angles at one or more joints and / or information about reaching specific relative positions of at least two joints and / or body parts to one another.

[0055] According to one embodiment, it is provided that status and / or movement-related data of the orthopaedic device are provided to the electronic computing unit, and the evaluation key figure is further determined as a function of the provided status and / or movement-related data of the orthopaedic device.

[0056] Such status and / or movement-related data of the orthopedic device can be, for example, joint angles of the orthopedic device, selected gait, and configurations of the orthopedic device, such as damping configurations. The object is also achieved according to the invention with the computer program according to claim 13. The computer program has program code means configured to carry out the method described above when the computer program is executed on an electronic data processing system.

[0057] The object is also achieved according to the invention with the device according to claim 14, in which the device has an electronic processing unit configured to carry out the method described above. The device can also contain a camera to record the video data of the movement sequence.

[0058] The invention is explained in more detail using the attached figure as an example. It shows:

[0059] Figure 1 Representation of the procedure;

[0060] Figure 2 Representation of a “timed up and go test”.

[0061] Figure 1 shows a simplified representation of the process flow of the present method for evaluating a movement sequence. First, a movement sequence is recorded using a camera 100 and provided in the form of digital video data 110.

[0062] This digital video data 110 is fed to a first machine learning system 120 for determining the body posture. The first machine learning system 120 receives the digital video data 110 as input and provides corresponding body features 130 on the patient's body contained in the digital video data 110 as output. These body features 130 are primarily joint points on the patient's body, wherein these joint points are detected and identified in some individual images, preferably in a majority, particularly preferably in all individual images of the digital video data 110.

[0063] The first machine learning system 120 was trained with training data containing a movement sequence and the body characteristics to be identified. The movement sequence can be, for example, an OMT, where the patient is fitted with an orthopedic device, such as a prosthesis or an orthosis. The prosthesis can be, for example, a leg prosthesis. Furthermore, the machine learning system 120 was preferably trained with various OMTs and patients with various orthopedic devices.

[0064] As an output, the machine learning system 120 can provide the two-dimensional coordinates within the individual image for each joint point 130 found in the individual image. The output of the first machine learning system 120 thus provides the position of the joint points in the entire video, thus enabling the determination or identification of a progression of the movement of these joint points in the video over time. Furthermore, other body features can be identified, such as the body parts connected at the joints, such as the thigh and lower leg.

[0065] The body characteristics 130 determined by the first machine learning system 120 are then provided to a calculation module 140, which identifies corresponding patient-related movement characteristics 150 from the previously determined body characteristics 130, in particular the joint points. Such patient-related movement characteristics 150 can, for example, be a joint angle and its progression over time. The distance between two joints and the change in such a distance over time can also be derived from the body characteristics 130, in particular the joint points. The touchdown point and the touchdown point when walking can also be determined, whereby for this purpose, the video data may be used if necessary, for example, to be able to recognize the ground.

[0066] These patient-related movement characteristics 150 are then fed to a second machine learning system 160. This second machine learning system 160 receives the previously determined patient-related movement characteristics 150 as input and provides a movement sequence type 170 as output, which characterizes the defined movement sequence performed by the patient in the video. With the help of the second machine learning system 160, the performed OMT is identified based on the patient-related movement characteristics 150.

[0067] Alternatively, the second machine learning system 160 can also be defined as a deterministic assignment rule set, which determines the movement sequence type 170 using a corresponding rule set based on the patient-related movement characteristics 150. This can be done, for example, via joint kinematics, in which a certain change in the joint angle over time indicates a certain movement sequence type with a certain probability.

[0068] The second machine learning system 160 is trained using training data, each of which contains a set of patient-related movement characteristics to which a specific movement sequence type is assigned. Through this assignment, the second machine learning system 160 can learn the assignment of these patient-related movement characteristics to a corresponding movement sequence type.

[0069] This movement sequence type 170 thus identified is provided, along with further information, such as the patient-related movement characteristics 150, to a verification module 180, which checks whether the movement sequence performed by the patient meets a predefined accuracy standard with respect to the assigned movement sequence type in terms of its execution. If this is not the case, an evaluation of the movement sequence cannot be performed. If this is the case, the movement sequence can be evaluated in the subsequent step using the evaluation module 200.

[0070] For example, the probability with which the second machine learning system 160 assigned a movement sequence to a specific type can be used to determine the correctness against the correctness measure. This probability can also be used to determine the evaluation score in the evaluation module 200.

[0071] Finally, in the evaluation module 200, the evaluation index 210 is determined. Among other things, the patient-related movement characteristics that were previously determined are used to determine, for example, movement information of the movement sequence and from this to determine the evaluation index 210. Such movement information can, for example, represent the distance traveled or the time for a certain distance, which can be derived from the video data and, in particular, the patient-related movement characteristics. In this process, movement information that is patient-related, such as uncertainties in gait or movement, can also be identified.

[0072] Figure 2 shows a schematic representation of the procedure for a "Timed Up and Go Test." Patient 300 initially sits on a chair. From there, they rise by standing up and then move a certain distance before returning to the chair and performing a sitting down movement. Throughout the entire test, the movement of patient 300, from the starting seated position to the final seated position, is recorded by a camera and evaluated according to the method described above.

[0073] For example, joint points are identified from which the knee angle can then be extracted as a movement feature.

[0074] In this example, the joint points are identified as body features. The knee joint 304, hip joint 302, and ankle joint 306 are shown. Alternatively, the body parts could also be identified directly, for example, the thigh and lower leg. The knee angle and its course could be considered as a movement feature here.

[0075] Examples for identifying the exercise: Is the knee angle initially approximately 90 degrees, and are the joint points of hip 302 and knee 304 approximately in a horizontal line? Does a reversal of movement occur approximately halfway through?

[0076] Examples to determine if the exercise was performed correctly: Was the armrest used to stand up (this would require identifying additional joint points). Was the stand-up performed in one go (no reversal of movement). Was the intended turning point reached?

[0077] Examples for evaluating the exercise: Duration of the exercise, where the beginning of standing up and the end of sitting down can be identified by the knee angle. If necessary, partial times can also be specified here, for example, until the reversal point of the movement or the time required to stand up or sit down. The maximum angular velocity when standing up is also an interesting parameter.

[0078] The evaluation using the present invention also allows the determination of additional parameters that the technician cannot, or can only with difficulty, measure manually. Thus, the present invention also solves one of the existing problems with these tests: namely, that the subtasks (standing up, turning around, etc.) cannot be analyzed individually; instead, it can only be determined whether an overall problem exists. A good result when standing up can therefore even mask a problem when walking, since the overall time remains within limits. The present invention detects such deviations.

[0079] List of reference symbols

[0080] 100 Camera

[0081] 110 digital video data 120 first machine learning system

[0082] 130 physical characteristics

[0083] 140 Calculation module

[0084] 150 patient-related movement characteristics

[0085] 160 second machine learning system 170 movement sequence type

[0086] 180 Verification module

[0087] 200 evaluation module

[0088] 210 Valuation key figure

[0089] 300 Patient 302 Hip joint

[0090] 304 Knee joint

[0091] 306 Ankle joint

Claims

Patent claims 1 . A method for evaluating a movement sequence performed by a patient (300) influenced by an orthopaedic device, the method comprising the following steps: - providing digital video data (110) containing a recording of the patient (300) influenced by the orthopaedic device during the execution of the predetermined movement sequence to an electronic computing unit, - Determining several patient-related movement characteristics (150) of the movement sequence of the patient (300) contained in the digital video data (110) by means of the electronic computing unit, - Assigning the movement sequence contained in the digital video data (110) to a movement sequence type (170) depending on the determined patient-related movement characteristics (150) by means of the electronic computing unit, - determining, by means of the electronic processing unit, whether the movement sequence contained in the digital video data (110) corresponds to a predefined degree of correctness with regard to the assigned movement sequence type (170) with regard to its execution, the determination being made as a function of the determined patient-related movement characteristics (150), and - Determining at least one evaluation index (210) as a function of the patient-related movement characteristics (150) by means of the computing unit.

2. Method according to one of the preceding claims, characterized in that a first machine learning system (120) is provided on the electronic computing unit, which system determines a relationship from movement sequences contained in digital video data (110) of patients (300) influenced by an orthopaedic device as input data and physical characteristics (130) of patients (300) as output data, wherein in order to determine physical characteristics (130) of the patient (300) in the movement sequence, the digital video data (110) are fed as input data to the first machine learning system (120) and the physical characteristics (130) of the patient (300) are then obtained as output data, wherein the patient-related movement characteristics (150) are identified as a function of the determined physical characteristics (130) of the patient (300).

3. The method according to claim 2, characterized in that the machine learning system (120) is an artificial neural network.

4. Method according to claim 2 or 3, characterized in that at least one joint point of a joint of the patient (300) in the movement sequence, a joint angle of at least one joint of the patient (300) in the movement sequence, a relative position between at least two joints of the patient (300) and / or a silhouette of the body of the patient (300) as well as changes in these body features (130) over time are identified as a body feature (130).

5. Method according to one of the preceding claims, characterized in that as patient-related movement characteristics (150) - a joint angle of a joint of the patient (300) at a specific point in time in the movement sequence or as a course over time, - a relative position between at least two joints of the patient (300) at a specific point in time in the movement sequence or as a course over time, and / or - a touchdown and / or a setdown point is determined in the movement sequence or as a progression over time.

6. Method according to one of the preceding claims, characterized in that a second machine learning system (160) is provided on the electronic computing unit, which system determines a relationship of patient-related movement characteristics (150) of the patient (300) when performing a predetermined movement sequence as input data and a movement sequence type (170) of the respective movement sequence as output data, wherein for the assignment of the movement sequence contained in the digital video data (110) to a corresponding movement sequence type (170), the determined patient-related movement characteristics (150) are fed to the second machine learning system (160) and the movement sequence type (170) is then obtained as output data.

7. Method according to one of the preceding claims, characterized in that an assignment rule set is provided on the electronic computing unit, which contains an assignment of certain combinations of patient-related movement characteristics (150) of the patient (300) when carrying out a predetermined movement sequence as input data to a respective movement sequence type (170) as output data, wherein for the assignment of the movement sequence contained in the digital video data (110) to a corresponding movement sequence type (170) the determined patient-related movement characteristics (150) are fed to the assignment rule set and the movement sequence type (170) is then obtained as output data.

8. Method according to claim 6 or 7, characterized in that the correctness measure corresponds to a probability in the assignment to the movement sequence type (170) or is derived therefrom.

9. Method according to one of the preceding claims, characterized in that the evaluation index (210) is determined by means of the computing unit as a function of a probability in the assignment to the movement sequence type (170).

10. Method according to one of the preceding claims, characterized in that the evaluation index (210) is determined as a function of the movement information of the movement sequence derived from the movement characteristics (150) is determined.

11. Method according to claim 10, characterized in that the movement information includes, among other things, the duration of the movement, the presence of a specific movement pattern, information about reaching specific joint angles at one or more joints and / or information about reaching specific relative positions of at least two joints and / or body parts to one another.

12. Method according to one of the preceding claims, characterized in that status and / or movement-related data of the orthopaedic device are provided to the electronic computing unit, the evaluation index (210) is further determined as a function of the provided status and / or movement-related data of the orthopaedic device.

13. Computer program with program code means configured to carry out the method according to one of claims 1 to 12 when the computer program is executed on an electronic data processing system.

14. Device with an electronic computing unit configured to carry out the method according to one of claims 1 to 12.

Images