Construction method and application of a cross-domain Parkinson's tremor assessment model

Abstract

The present invention discloses a method for constructing a cross-domain Parkinson's tremor assessment model and its application, belonging to the field of artificial intelligence technology. By minimizing the total tremor classification loss and maximizing the total domain classification loss, a feature extraction module, a tremor classifier, a domain classifier and a weight allocator are trained simultaneously. After the training is completed, a cross-domain Parkinson's tremor assessment model including a cascaded feature extraction module and a tremor classifier is constructed. The total domain classification loss is the sum of the domain classification losses of the source domain and the target domain. When calculating the domain classification losses of different domains, after assigning weights to each training sample under each domain through the weight allocator, the domain classification losses of each training sample obtained based on the domain classifier are weighted at the sample level. This fully utilizes the differences between Parkinson's tremor sensor data, promotes the alignment of the source domain and target domain feature spaces, and enables the constructed model to accurately assess Parkinson's tremor.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and more specifically, relates to a method for constructing a cross-domain Parkinson's tremor assessment model and its application. Background Art

[0002] Parkinson's disease is a common progressive neurodegenerative disease. Tremor is one of the common symptoms caused by Parkinson's disease, and its severity usually indicates the progression of the disease. The clinical assessment method for Parkinson's tremor is usually based on the MDS Unified Parkinson's Disease Rating Scale (MDS-UPDRS), which divides it into integers of 0-4 levels (0 for normal and 4 for severe) according to the manifestation of the disease. However, this assessment method has disadvantages such as subjectivity, time-consuming and insufficient comprehensiveness, resulting in limited reliability of the assessment results.

[0003] In recent years, traditional machine learning and deep learning methods have been widely used for Parkinson's tremor assessment. Regardless of which method is used, relatively good evaluation results can be achieved if user-specific data partitioning is used, i.e., each subject's data is divided into a training set and a test set. However, accuracy decreases significantly if user-independent data partitioning is used, i.e., data from the same subject does not appear in both the training and test sets. This phenomenon is primarily due to the fact that traditional machine learning and deep learning methods require that the training and test sets adhere to the independent and identically distributed (IID) assumption. These partitioning methods ignore the influence of data distribution differences between different users. In reality, Parkinson's tremor data is affected by various factors, such as age, gender, and data collection environment. Therefore, the IID assumption is often not met for both training and test sets, resulting in poor model generalization. Furthermore, these approaches are all based on supervised learning methods, which require the collection and labeling of sufficient data for model training. However, obtaining large-scale, high-precision data labels is a significant challenge in Parkinson's tremor research. There are two main reasons for this. First, tremor data requires annotation by specialists with extensive clinical experience, a process that is time-consuming, labor-intensive, and subjective. Second, experimental data contains patient privacy, and privacy regulations make it difficult to share experimental data across different research centers. To address these issues, domain adaptation methods have been proposed. They transfer knowledge learned from the source domain to the target domain, creating domain-invariant feature representations. Currently, mainstream methods can be divided into two categories: one is domain adaptation methods based on distance metrics, such as Maximum Mean Discrepancy (MMD) and KL Divergence; the other is adversarial domain adaptation methods, such as DANN (Domain-Adversarial Neural Network) and CDAN (Conditional Adversarial Domain Adaptation). However, these methods are currently primarily applied to image datasets and time series classification tasks, with relatively little research on Parkinson's tremor assessment. Furthermore, these methods often indiscriminately align samples from different domains during the adaptation phase, failing to fully consider the differences between samples. In Parkinson's tremor sensor data, even within the same tremor score level, there are significant differences between different users. If a user's tremor data contains significantly more personalized information than the tremor characteristics themselves, then that user's data may negatively impact the adaptation process. Consequently, existing domain adaptation methods have limitations in accurately assessing Parkinson's tremor. Summary of the Invention

[0004] In response to the above-mentioned defects or improvement needs of the existing technology, the present invention provides a method for constructing a cross-domain Parkinson's tremor assessment model and its application, so as to solve the technical problem that the existing technology cannot accurately assess Parkinson's tremor.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for constructing a cross-domain Parkinson's tremor assessment model, comprising:

[0006] The Parkinson's tremor sensor data collected in two different environments are used as source domain training samples and target domain training samples respectively, and then a source domain training sample set with tremor score level labels and a target domain training sample set without tremor score level labels are constructed;

[0007] Input multiple pairs of training samples for training in each training batch:

[0008] Each time a pair of source domain training samples x s and target domain training samples x T Input into the feature extraction module and extract x s 、x T The sample feature F s 、F T ; F s and F T Input them into the tremor classifier respectively to get x s 、x T The tremor score prediction results are calculated and x s 、x T Tremor classification loss; x s The tremor classification loss is used to measure x s The difference between the tremor score grade prediction result and the corresponding tremor score grade label; T The tremor classification loss is used to measure x T The difference between the predicted tremor score level and the corresponding tremor score level pseudo label;

[0009] F s and F T Input into the domain classifier respectively to get x s and x T The domain prediction result of , and calculate x s 、x T Domain classification loss; domain classification loss is used to measure the difference between the predicted result of the domain and the corresponding true result; respectively for x s and x T , concatenate its domain classification loss and tremor classification loss to obtain the corresponding concatenation loss;

[0010] For each domain, the concatenation loss of all the training samples of the corresponding domain input in the current training batch is input into the weight distributor. After obtaining the weight value of each training sample of the corresponding domain, the domain classification loss of each training sample of the corresponding domain is weighted and summed to obtain the domain classification loss of the corresponding domain.

[0011] Calculate the sum of the domain classification losses of the source domain and the target domain to obtain the total domain classification loss;

[0012] Calculate the sum of the tremor classification losses of all input source domain training samples and target domain training samples to obtain the total tremor classification loss;

[0013] The feature extraction module, tremor classifier, domain classifier, and weight allocator are trained simultaneously by minimizing the total tremor classification loss and maximizing the total domain classification loss.

[0014] After training, a cross-domain Parkinson's tremor assessment model consisting of a cascaded feature extraction module and a tremor classifier was constructed.

[0015] More preferably, x s The tremor classification loss is: s The difference loss between the predicted probability vector and the corresponding true probability vector;

[0016] Among them, x s The predicted probability vector is obtained by s Input to the tremor classifier to get; x s The i-th probability value in the predicted probability vector is x s The probability of being predicted as the i-th tremor score level; i = 1, 2, ..., n; n is the number of tremor score levels; x s The predicted result of tremor score is x s The tremor score level corresponding to the maximum probability value in the predicted probability vector;

[0017] The true probability vector is a vector composed of n probability values, in which the idx-th probability value is 1 and the rest are 0; idx is x s The index of the tremor score grade label.

[0018] More preferably, when x T The predicted probability vector P T When the maximum probability value in is greater than or equal to the preset threshold ρ, x T The tremor classification loss is: P T The difference loss with the corresponding pseudo probability vector; where P T By adding F T Input to the tremor classifier to obtain; P T The i-th probability value in is x TThe probability of being predicted as the i-th tremor score level; x T The tremor score prediction result is P T The tremor score level corresponding to the maximum probability value in ; the pseudo probability vector is obtained by T The maximum value in is set to 1, and the rest are set to 0;

[0019] When P T When the maximum probability value in is less than the preset threshold ρ, x T The tremor classification loss is 0.

[0020] Further preferably, the feature extraction module includes: a time domain feature extractor, a frequency domain feature extractor and a fusion module;

[0021] The time domain feature extractor is used to extract the time domain features of the input samples;

[0022] The frequency domain feature extractor is used to extract the frequency domain features of the frequency domain signal after the input sample undergoes time-frequency conversion;

[0023] The fusion module is used to fuse the time domain features and frequency domain features to obtain sample features.

[0024] Further preferably, the time domain feature extractor and the frequency domain feature extractor are both CNN models.

[0025] Further preferably, the time domain feature extractor and the frequency domain feature extractor both include: a cascaded convolution module and a global average pooling layer;

[0026] The convolution module includes: a cascaded one-dimensional convolution layer, a batch normalization layer, and an activation function layer.

[0027] In a second aspect, the present invention provides a cross-domain Parkinson's tremor assessment method, comprising: inputting collected tremor sensor data into a cross-domain Parkinson's tremor assessment model to obtain a corresponding tremor score level;

[0028] The cross-domain Parkinson's tremor assessment model is constructed using the method for constructing a cross-domain Parkinson's tremor assessment model provided by the first aspect of the present invention.

[0029] In a third aspect, the present invention further provides a computer-readable storage medium, comprising a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the method for constructing a cross-domain Parkinson's tremor assessment model provided in the first aspect of the present invention or the cross-domain Parkinson's tremor assessment method provided in the second aspect of the present invention.

[0030] In a fourth aspect, the invention also provides a computer program product, comprising a computer program / instruction, characterized in that when the computer program / instruction is executed by a processor, it implements the method for constructing a cross-domain Parkinson's tremor assessment model provided in the first aspect of the invention or the cross-domain Parkinson's tremor assessment method provided in the second aspect of the invention.

[0031] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0032] 1. The present invention provides a method for constructing a cross-domain Parkinson's tremor assessment model, which minimizes the total tremor classification loss and maximizes the total domain classification loss while simultaneously training a feature extraction module, a tremor classifier, a domain classifier, and a weight allocator. After the training is completed, a cross-domain Parkinson's tremor assessment model including a cascaded feature extraction module and a tremor classifier is constructed. The total domain classification loss is the sum of the domain classification losses of the source domain and the target domain. When calculating the domain classification losses of different domains, after assigning weights to each training sample under each domain through the weight allocator, the domain classification losses of each training sample obtained based on the domain classifier are weighted at the sample level, thereby making full use of the differences between Parkinson's tremor sensor data, promoting the alignment of the feature spaces of the source domain and the target domain, and enabling the constructed model to accurately assess Parkinson's tremor.

[0033] 2. Furthermore, the method for constructing a cross-domain Parkinson's tremor assessment model provided by the present invention sets a threshold ρ for the classification confidence of the target domain training samples in order to make the model focus more on reliable predictions during the training process. The target domain training samples with a confidence level higher than the threshold ρ are selected to calculate the classification loss, thereby further improving the accuracy of Parkinson's tremor assessment.

[0034] 3. Furthermore, the method for constructing a cross-domain Parkinson's tremor assessment model provided by the present invention extracts features in both the time domain and frequency domain modes for Parkinson's tremor sensor data, and then constructs corresponding multimodal features. This not only fully utilizes the correlation in the time domain, but also fully utilizes the information in the frequency domain, further improving the accuracy of Parkinson's tremor assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of a process for constructing a cross-domain Parkinson's tremor assessment model provided by an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of the structure of a feature extraction module provided in an embodiment of the present invention;

[0037] Figure 3 A schematic diagram of feature splicing in the column direction of time domain features and frequency domain features provided in an embodiment of the present invention;

[0038] Figure 4 A schematic structural diagram of a tremor classifier provided in an embodiment of the present invention;

[0039] Figure 5 A schematic diagram of the structure of a weight allocator provided in an embodiment of the present invention;

[0040] Figure 6 A schematic diagram of the structure of a domain classifier provided in an embodiment of the present invention;

[0041] Figure 7 A framework diagram of a cross-domain Parkinson's tremor assessment model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0043] To achieve the above objectives, in a first aspect, the present invention provides a method for constructing a cross-domain Parkinson's tremor assessment model, comprising:

[0044] S1. Use Parkinson's tremor sensor data collected in two different environments as source domain training samples and target domain training samples, respectively, to construct a source domain training sample set with tremor score level labels and a target domain training sample set without tremor score level labels; the number of samples in the source domain training sample set is smaller than the number of samples in the target domain training sample set;

[0045] S2. Input multiple pairs of training samples for training in each training batch:

[0046] S21, each time a pair of source domain training samples x s and target domain training samples x T Input into the feature extraction module and extract x s 、x T The sample feature F s 、F T ; F s and F T Input them into the tremor classifier respectively to get x s 、x T The tremor score prediction results are calculated and x s 、x T Tremor classification loss; x s The tremor classification loss is used to measure x sThe difference between the tremor score grade prediction result and the corresponding tremor score grade label; T The tremor classification loss is used to measure x T The difference between the predicted tremor score level and the corresponding tremor score level pseudo label;

[0047] It should be noted that the feature extraction module can be CNN, RNN, LSTM, Transformer, etc., which is not limited here, and a fully connected layer is preferably used.

[0048] Preferably, in an optional implementation manner, the feature extraction module includes: a time domain feature extractor, a frequency domain feature extractor and a fusion module;

[0049] The time domain feature extractor is used to extract the time domain features of the input samples;

[0050] The frequency domain feature extractor is used to extract the frequency domain features of the frequency domain signal after the input sample undergoes time-frequency conversion;

[0051] The fusion module is used to fuse time-domain features and frequency-domain features to obtain sample features. It should be noted that there are many ways to fuse features, including column-wise concatenation, direct addition of corresponding elements, weighted merging based on feature relevance or importance, and principal component analysis.

[0052] It should be noted that the time domain feature extractor and the frequency domain feature extractor are independent of each other and do not share weight parameters.

[0053] In an optional implementation, both the time domain feature extractor and the frequency domain feature extractor are CNN models.

[0054] In an optional embodiment, the time domain feature extractor and the frequency domain feature extractor each include: a cascaded convolution module and a global average pooling layer;

[0055] The convolution module includes: a cascaded one-dimensional convolution layer, a batch normalization layer, and an activation function layer.

[0056] In one embodiment, the tremor classifier is used to obtain the probability of a training sample being predicted as having different tremor score levels, and the tremor score level corresponding to the maximum probability is used as the tremor score level prediction result of the training sample. The tremor score level pseudo label is obtained based on the predicted probability of the tremor classifier. Specifically, in one optional embodiment, x s The tremor classification loss is: s The difference loss between the predicted probability vector and the corresponding true probability vector;

[0057] Among them, x s The predicted probability vector is obtained by sInput to the tremor classifier to get; x s The i-th probability value in the predicted probability vector is x s The probability of being predicted as the i-th tremor score level; i = 1, 2, ..., n; n is the number of tremor score levels; x s The predicted result of tremor score is x s The tremor score level corresponding to the maximum probability value in the predicted probability vector;

[0058] The true probability vector is a vector composed of n probability values, in which the idx-th probability value is 1 and the rest are 0; idx is x s The index of the tremor score grade label.

[0059] In an alternative embodiment, when x T The predicted probability vector P T When the maximum probability value in is greater than or equal to the preset threshold ρ, x T The tremor classification loss is: P T The difference loss with the corresponding pseudo probability vector; where P T By adding F T Input to the tremor classifier to obtain; P T The i-th probability value in is x T The probability of being predicted as the i-th tremor score level; x T The tremor score prediction result is P T The tremor score level corresponding to the maximum probability value in ; the pseudo probability vector is obtained by T The maximum value in is set to 1, and the rest are set to 0;

[0060] When P T When the maximum probability value in is less than the preset threshold ρ, x T The tremor classification loss is 0.

[0061] In another embodiment, pseudo labels for tremor rating levels of target domain training samples are obtained by classifying source domain training samples according to tremor rating level labels and calculating the average of features extracted by a feature extraction module for source domain training samples at different tremor rating levels as reference features for the corresponding tremor rating levels. Features of the target domain training samples are obtained by the feature extraction module, and similarities are calculated between the features of the target domain training samples and the reference features at different tremor rating levels. The tremor rating level corresponding to the reference feature with the smallest similarity is used as the pseudo label for the tremor rating level of the target domain training sample.

[0062] It should be noted that the tremor classifier can be a traditional machine learning model (such as SVM, logistic regression, decision tree, etc.) and a deep learning model (such as fully connected neural network, CNN, LSTM, etc.), etc., which is not limited here, and a fully connected neural network is preferably used.

[0063] S22, F s and F T Input into the domain classifier respectively to get x s and x T The domain prediction result of , and calculate x s 、x T Domain classification loss; domain classification loss is used to measure the difference between the predicted result of the domain and the corresponding true result; respectively for x s and x T , concatenate its domain classification loss and tremor classification loss to obtain the corresponding concatenation loss;

[0064] It should be noted that the domain classifier can be a traditional machine learning model (such as SVM, logistic regression, decision tree, etc.) and a deep learning model (such as fully connected neural network, CNN, LSTM, etc.), etc., which is not limited here. A fully connected neural network is preferably used.

[0065] S23. For each domain, the concatenation loss of all training samples of the corresponding domain input in the current training batch is input into the weight distributor. After obtaining the weight value of each training sample of the corresponding domain, the domain classification loss of each training sample of the corresponding domain is weighted and summed to obtain the domain classification loss of the corresponding domain; specifically, the concatenation loss of all source domain training samples in the current training batch is input into the weight distributor. After obtaining the weight value of each source domain training sample in the current training batch, the domain classification loss of each source domain training sample is weighted and summed to obtain the domain classification loss of the source domain; the concatenation loss of all target domain training samples in the current training batch is input into the weight distributor. After obtaining the weight value of each target domain training sample in the current training batch, the domain classification loss of each target domain training sample is weighted and summed to obtain the domain classification loss of the target domain; the sum of the total classification losses of the source domain and the target domain is calculated to obtain the total domain classification loss;

[0066] Calculate the sum of the tremor classification losses of all input source domain training samples and target domain training samples to obtain the total tremor classification loss;

[0067] It should be noted that the weight allocator can be a fully connected neural network, CNN, RNN, LSTM, attention network, etc., which is not limited here, and a fully connected neural network is preferably used.

[0068] S24, simultaneously training the feature extraction module, the tremor classifier, the domain classifier, and the weight allocator by minimizing the total tremor classification loss and maximizing the total domain classification loss;

[0069] S3. After training is completed, a cross-domain Parkinson's tremor assessment model including a cascaded feature extraction module and a tremor classifier is constructed.

[0070] To further illustrate the method for constructing a cross-domain Parkinson's tremor assessment model provided by the present invention, a detailed description is given below in conjunction with a specific embodiment:

[0071] This embodiment uses wearable sensor-based cross-domain Parkinson's tremor assessment as an example to illustrate. The signals used for Parkinson's tremor acquisition based on wearable sensors are generally acceleration signals and gyroscope signals. These sensors are fixed to various body parts of the subject, such as the wrist, calf, and foot, to detect the occurrence of tremor in the subject. Parkinson's tremor scoring is currently mainly based on the MDS-UPDRS scale standard in clinical practice. The integer is divided into 0-4 levels (0 represents normal and 4 represents severe) based on the symptoms during outpatient examinations. The sub-items related to tremor in this scale are 3.15 to 3.18, which are respectively the postural tremor of the hand, the action tremor of the hand, the amplitude of the resting tremor, and the persistence of the resting tremor. Postural tremor, action tremor, and resting tremor are all Parkinson's tremors. Different tremor types have different characteristics, but the invention is applicable to all analyses involving tremor behavior itself. As a specific embodiment, take the upper limb resting tremor data collected in a controlled environment as an example. For the resting tremor task, participants were asked to sit in a chair with their arms resting on the armrests, and a measurement device was attached to each participant's wrist. Because tremors are not continuous when they occur, label reviewers carefully reviewed the recordings with the aid of synchronized video and recorded the timestamps corresponding to the onset and end of each patient's visible tremor. This embodiment of the present invention has no restrictions on the signal type; therefore, commonly used acceleration and gyroscope signals, as well as other relevant multimodal signals, can be used as network inputs.

[0072] like Figure 1 As shown in Figure 2, the construction method of the cross-domain Parkinson's tremor assessment model includes:

[0073] 1) Data acquisition and preprocessing. After obtaining the informed consent of the subjects to participate in the experiment, the tremor sensor data and the corresponding MDS-UPDRS tremor subscore of the subjects were collected according to a certain activity protocol. The tremor data were preprocessed by filling missing values, filtering, etc.; the tremor signal X t Perform time-frequency mode conversion to obtain the frequency domain signal X f According to the cross-domain scenario set in the experiment, the acquired data is divided into labeled source domains and the unlabeled target domain Further sliding window segmentation to construct source domain training samples Target domain training samples Target domain test samples

[0074] During signal preprocessing, the sensor data were first timestamp aligned and downsampled to a common frequency, set at 100 Hz. The signals were then checked for duplicate rows and missing values, with duplicate rows removed and missing values ​​filled using linear interpolation. The signals were then subjected to a Butterworth bandpass filter to extract the typical frequencies of the corresponding tremor type, such as 3-7 Hz for resting tremor. Low- and high-frequency noise and voluntary movements were also filtered out, preserving only the tremor component as much as possible. To simultaneously capture both the time and frequency domains of the tremor signal, a one-dimensional multiscale discrete wavelet transform was used to convert the time-domain signal into the frequency domain.

[0075] For sample division, first divide the samples into source domain and target domain according to the cross-domain scenario set in the experiment. The cross-domain scenario can be any factor that causes inconsistent data distribution between the Parkinson's tremor source domain and the target domain, such as differences in gender, age, and data collection environment. If the cross-domain scenario of the experiment is set to different test centers, the subject data from test center 1 will be used as the source domain data, and the subject data from test center 2 will be used as the target domain data. All data in the source domain will be used as source domain training samples, and the target domain data will be divided into target domain training samples and target domain test samples in a 1:1 ratio. The actual ratio can be adjusted according to the number of samples.

[0076] 2) The source domain training samples and target domain training samples The time domain signal is input to the time domain feature extractor F t In the process, we obtain the time domain feature fea T ; The source domain training samples and target domain training samples The frequency domain signal is input to the frequency domain feature extractor F f In the frequency domain, obtain the fea F .

[0077] In this embodiment, a parallel deep network is used to extract features from the time domain and frequency domain of the signal. Figure 2 In an optional embodiment, both the time domain feature extractor and the frequency domain feature extractor use convolutional neural networks. The network structure includes three convolution blocks and a global average pooling layer. Each convolution block includes: a one-dimensional convolution layer, a batch normalization layer, and an activation function layer. The specific structure is as follows:

[0078] The first one-dimensional convolution block includes: a convolution layer with 64 convolution kernels, a size of 5 and a stride of 2, a batch normalization layer and a ReLU activation function layer;

[0079] The second one-dimensional convolution block includes: a convolution layer with 128 convolution kernels, a size of 3 and a stride of 2, a batch normalization layer and a ReLU activation function layer;

[0080] The third one-dimensional convolution block includes: a convolution layer with 64 convolution kernels, a size of 2 and a stride of 1, a batch normalization layer and a ReLU activation function layer.

[0081] 3) The acquired time domain features fea T and frequency domain features F The multimodal features (sample features) fea are directly concatenated and then input into the tremor classifier C and the domain classifier D respectively to obtain the classification probability and domain classification probability.

[0082] like Figure 3 As shown in the figure, the acquired time domain features and frequency domain features are concatenated in the column direction, and the concatenated features are called multimodal features.

[0083] In this embodiment, the structure of the tremor classifier is a fully connected layer and an activation function layer, and the output dimension is the number of tremor assessment levels. If the labels of the training samples cover all levels of the MDS-UPDRS tremor score, the output dimension should be set to 5. When the activation function is the softmax function, the tremor classifier outputs the classification probability That is, the probability that the sample belongs to each tremor level, n c is the number of categories. The structure of the domain classifier is fully connected layer 1, ReLU layer, fully connected layer 2, ReLU layer, fully connected layer 3, Sigmoid layer, and its output is the domain classification probability.

[0084] 4) Calculate the total tremor classification loss. The total tremor classification loss consists of two parts: the domain classification loss of the source domain and the domain classification loss of the target domain. Since the target domain training samples do not contain label information, pseudo labels are used for calculation. Taking the target domain training samples as input, the tremor classifier outputs the tremor level corresponding to the maximum probability in the classification probability, which is the pseudo label, and the corresponding maximum probability value is the classification confidence. In order to make the model pay more attention to reliable predictions during the training process, a threshold ρ is set for the classification confidence of the target domain training samples, and the target domain training samples above the threshold ρ are selected to calculate the classification loss. In this embodiment, ρ is taken as 0.7.

[0085] Specifically, the total tremor classification loss of the source domain is the sum of the tremor classification losses of each source domain training sample; s The tremor classification loss is: the source domain training sample xs The difference loss between the predicted probability vector and the corresponding true probability vector;

[0086] Among them, x s The predicted probability vector is obtained by s Input to the tremor classifier to get; x s The i-th probability value in the predicted probability vector is x s The probability of being predicted as the i-th tremor score level; i = 1, 2, ..., n; n is the number of tremor score levels; x s The predicted result of tremor score is x s The tremor score level corresponding to the maximum probability value in the predicted probability vector;

[0087] The true probability vector is a vector composed of n probability values, in which the idx-th probability value is 1 and the rest are 0; idx is x s The index of the tremor score grade label.

[0088] The total tremor classification loss of the target domain is the sum of the tremor classification losses of each target domain training sample; T The predicted probability vector P T When the maximum probability value in is greater than or equal to the preset threshold ρ, the target domain training sample x T The tremor classification loss is: P T The difference loss with the corresponding pseudo probability vector; where P T By adding F T Input to the tremor classifier to obtain; P T The i-th probability value in is x T The probability of being predicted as the i-th tremor score level; x T The tremor score prediction result is P T The tremor score level corresponding to the maximum probability value in ; the pseudo probability vector is obtained by T The maximum value in is set to 1, and the rest are set to 0; when P T When the maximum probability value in is less than the preset threshold ρ, x T The tremor classification loss is 0.

[0089] 5) The tremor classification loss and domain classification loss of each source domain training sample are directly spliced ​​and input into the weight allocator. The output is normalized to be the weight of each source domain training sample. Each source domain training sample is multiplied by the corresponding weight and then added together to obtain the domain classification loss of the source domain.

[0090] The tremor classification loss and domain classification loss of each target domain training sample are directly concatenated and input into the weight distributor. The output is normalized to be the weight of each target domain training sample. Each target domain training sample is multiplied by the corresponding weight and then added together to obtain the domain classification loss of the target domain.

[0091] The domain classification loss of the source domain and the domain classification loss of the target domain are summed to obtain the total domain classification loss.

[0092] In this embodiment, the weight allocator includes: a cascaded first fully connected layer, a ReLU layer, a second fully connected layer, and a Sigmoid layer.

[0093] 6) Source domain training samples and target domain training samples This example uses adversarial training and dual optimization to minimize the total tremor classification loss while maximizing the total domain classification loss. This makes it difficult for the domain classifier to classify features from the source and target domains, thereby extracting domain-invariant features. The specific training steps are divided into the following three steps:

[0094] The first step is outer loop optimization. Use source domain training data and target domain training data Train the feature extraction module F and the tremor classifier C. The training data is output by the feature extraction module and the tremor classifier to obtain the corresponding classification probability. Since the target domain does not contain true labels, pseudo labels are used to calculate the classification loss. The loss in this step is the total tremor classification loss.

[0095] The second step is inner loop optimization. Use source domain training samples and target domain training samples To train the weight allocator W, we first create a replica of the feature extraction module from the first step. After the replica extracts multimodal features from the training data, it is fed into the tremor classifier and domain classifier, respectively. The tremor and domain classification losses are calculated for each training example. The tremor and domain classification losses for each source domain training example are then concatenated to obtain the corresponding concatenation loss. The tremor and domain classification losses for each target domain training example are also concatenated to obtain the corresponding concatenation loss. The splicing loss of each source domain training sample is input into the weight distributor to obtain the sample weight, and the domain classification loss of each source domain training sample is weightedly summed to obtain the total domain classification loss of the source domain; the splicing loss of each target domain training sample is input into the weight distributor to obtain the sample weight, and the domain classification loss of each target domain training sample is weightedly summed to obtain the total domain classification loss of the target domain; the sum of the total domain classification loss of the source domain and the total domain classification loss of the target domain is calculated to obtain the total domain classification loss; by maximizing the total domain classification loss, the parameters of the feature extraction module copy are updated; on this basis, the tremor classification loss of each target domain training sample is calculated and summed to obtain the total tremor classification loss under the target domain; the loss is optimized to update the parameters of the weight distributor.

[0096] In this embodiment, maximizing the total domain classification loss is achieved through a gradient reversal layer. The gradient reversal layer can achieve an identity transformation of the gradient during forward propagation and automatically negate it during backward propagation.

[0097]

[0098] Where λ is a constant and I is the identity matrix.

[0099] The third step is outer loop optimization. Use source domain training samples and target domain training samples Train the feature extraction module and domain classifier, calculate the total domain classification loss, and optimize the loss to update the parameters of the feature extraction module and domain classifier.

[0100] In this embodiment, the tremor classification loss is calculated based on the cross entropy loss function, and the domain classification loss is calculated based on the binary cross entropy loss function.

[0101] In this embodiment, the structure of the tremor classifier is a fully connected layer with 5 output neurons and a Sigmoid layer, such as Figure 4 The structure of the weight distributor is a linear layer with 128 output neurons, a ReLU layer, a linear layer with 2 output neurons, and a Sigmoid layer, as shown in Figure 5 As shown in Figure 2, the structure of the domain classifier is linear layer 1, batch normalization layer, ReLU layer, linear layer 2, batch normalization layer, ReLU layer, linear layer 3 and Sigmoid layer, as shown in Figure 2. Figure 6 As shown in Figure 2, the number of output neurons in the three linear layers are 256, 256, and 2 respectively.

[0102] In this example, the network was trained using the Adam optimizer. The initial learning rate for all subnetworks was set to 0.001, and the cosine annealing algorithm was used for learning rate adjustment. The batch size was set to 128 and the maximum number of iterations, n_epoch, was set to 10 for both training and testing. Training was terminated when the set number of iterations was reached, resulting in an optimal tremor classifier. The inner loop was optimized using a differentiable optimizer provided by Higher.

[0103] Considering the imbalanced nature of the tremor dataset, this example adds the Kappa coefficient and MCC in addition to the three commonly used evaluation metrics—accuracy, sensitivity, and F1 index. The Kappa coefficient provides a more realistic assessment than accuracy because it penalizes model bias. MCC is a comprehensive classification performance evaluation metric that mitigates classifier bias caused by uneven sample distribution.

[0104] Finally, the target domain test samples are classified. After these steps, the network has completed the transfer of target domain knowledge. The target domain test sample data is input, and the trained multimodal feature extractor and tremor classifier are used to obtain the final corresponding tremor score prediction results, thus evaluating the model classification performance.

[0105] The entire training process is as follows Figure 7 shown.

[0106] In a second aspect, the present invention provides a cross-domain Parkinson's tremor assessment method, comprising: inputting collected tremor sensor data into a cross-domain Parkinson's tremor assessment model to obtain a corresponding tremor score level;

[0107] The cross-domain Parkinson's tremor assessment model is constructed using the method for constructing a cross-domain Parkinson's tremor assessment model provided by the first aspect of the present invention;

[0108] Specifically, the collected tremor sensor data is input into the cross-domain Parkinson's tremor assessment model. After feature extraction by the feature extraction module, the tremor sensor data is input into the tremor classifier to obtain the corresponding tremor score level.

[0109] The related technical solutions are the same as the method for constructing a cross-domain Parkinson's tremor assessment model provided by the first aspect of the present invention, and will not be described in detail here.

[0110] In a third aspect, the present invention further provides a computer-readable storage medium, comprising a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the method for constructing a cross-domain Parkinson's tremor assessment model provided in the first aspect of the present invention or the cross-domain Parkinson's tremor assessment method provided in the second aspect of the present invention.

[0111] The relevant technical solutions are the same as the method for constructing a cross-domain Parkinson's tremor assessment model provided by the first aspect of the present invention and the cross-domain Parkinson's tremor assessment method provided by the second aspect of the present invention, and will not be described in detail here.

[0112] In a fourth aspect, the invention also provides a computer program product, comprising a computer program / instruction, characterized in that when the computer program / instruction is executed by a processor, it implements the method for constructing a cross-domain Parkinson's tremor assessment model provided in the first aspect of the invention or the cross-domain Parkinson's tremor assessment method provided in the second aspect of the invention.

[0113] The relevant technical solutions are the same as the method for constructing a cross-domain Parkinson's tremor assessment model provided by the first aspect of the present invention and the cross-domain Parkinson's tremor assessment method provided by the second aspect of the present invention, and will not be described in detail here.

[0114] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a cross-domain Parkinson's tremor assessment model, characterized in that: include: The Parkinson's tremor sensor data collected in two different environments were used as source domain training samples and target domain training samples, respectively, to construct a source domain training sample set with tremor score level labels and a target domain training sample set without tremor score level labels. Input multiple pairs of training samples for training in each training batch: Each time a pair of source domain training samples and target domain training samples Input into the feature extraction module and extract 、 Sample characteristics 、 ;Will and Input them into the tremor classifier respectively, and get 、 The tremor score prediction results are calculated and 、 tremor classification loss; The tremor classification loss is used to measure The difference between the predicted tremor score level and the corresponding tremor score level label; The tremor classification loss is used to measure The difference between the predicted tremor score level and the corresponding tremor score level pseudo label; Will and Input them into the domain classifier respectively, and get and The domain prediction results of 、 Domain classification loss; Domain classification loss is used to measure the difference between the predicted result of the domain and the corresponding true result; and , concatenate its domain classification loss and tremor classification loss to obtain the corresponding concatenation loss; For each domain, the concatenation loss of all the training samples of the corresponding domain input in the current training batch is input into the weight distributor. After obtaining the weight value of each training sample of the corresponding domain, the domain classification loss of each training sample of the corresponding domain is weighted and summed to obtain the domain classification loss of the corresponding domain. Calculate the sum of the domain classification losses of the source domain and the target domain to obtain the total domain classification loss; Calculate the sum of the tremor classification losses of all input source domain training samples and target domain training samples to obtain the total tremor classification loss; The feature extraction module, tremor classifier, domain classifier, and weight allocator are trained simultaneously by minimizing the total tremor classification loss and maximizing the total domain classification loss. After training, a cross-domain Parkinson's tremor assessment model consisting of a cascaded feature extraction module and a tremor classifier was constructed; The feature extraction module includes: a time domain feature extractor, a frequency domain feature extractor and a fusion module; The time domain feature extractor is used to extract the time domain features of the input samples; The frequency domain feature extractor is used to extract the frequency domain features of the frequency domain signal after the input sample undergoes time-frequency conversion; The fusion module is used to fuse the time domain features and the frequency domain features to obtain sample features.

2. The method for constructing a cross-domain Parkinson's tremor assessment model according to claim 1, characterized in that: described The tremor classification loss is: The difference loss between the predicted probability vector and the corresponding true probability vector; Among them, the The predicted probability vector is calculated by Input into the tremor classifier to obtain; The first in the predicted probability vector i The probability value is Predicted to be i probability of achieving a tremor score level; ; n the number of grades scored for tremor; The predicted tremor score is The tremor score level corresponding to the maximum probability value in the predicted probability vector; The true probability vector is n The vector of probability values, idx The probability value is 1, and the rest are 0; idx for The index of the tremor score grade label.

3. The method for constructing a cross-domain Parkinson's tremor assessment model according to claim 1, characterized in that: when The predicted probability vector The maximum probability value in is greater than or equal to the preset threshold When The tremor classification loss is: The difference loss with the corresponding pseudo probability vector; where, By Input to the tremor classifier to obtain; The i The probability value is Predicted to be i probability of achieving a tremor score level; The predicted tremor score is The tremor score level corresponding to the maximum probability value in The maximum value in is set to 1, and the rest are set to 0; when The maximum probability value in is less than the preset threshold When The tremor classification loss is 0.

4. The method for constructing a cross-domain Parkinson's tremor assessment model according to claim 1, characterized in that: The time domain feature extractor and the frequency domain feature extractor are both CNN models.

5. The method for constructing a cross-domain Parkinson's tremor assessment model according to claim 4, characterized in that: The time domain feature extractor and the frequency domain feature extractor both include: a cascaded convolution module and a global average pooling layer; The convolution module includes: a cascaded one-dimensional convolution layer, a batch normalization layer and an activation function layer.

6. A cross-domain Parkinson's tremor assessment method, characterized in that: include: The collected tremor sensor data is input into the cross-domain Parkinson's tremor assessment model to obtain the corresponding tremor score level; The cross-domain Parkinson's tremor assessment model is constructed using the method for constructing a cross-domain Parkinson's tremor assessment model according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed by a processor, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

8. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images