29 results about "Autism child" patented technology

The invention discloses a child autism adaptive brain-computer fusion intervention system based on a large model, and the system comprises a multi-modal data collection module which is used for collecting the neural data, behavior data and clinical scale data of a user; the feature extraction and user portrait construction module is used for performing feature extraction and subtype recognition on the multi-modal data to generate a personalized portrait; the multi-defect intervention normal form generation module is used for dynamically generating a personalized intervention task on the basis of a large language model in combination with the personalized portrait and the historical state of the task; the self-adaptive regulation and control module is used for dynamically adjusting an intervention strategy and task difficulty according to the feedback of the real-time neural data and the dynamic change of the behavior data; the user interaction interface module is used for providing a multi-modal human-computer interaction interface; and the effect evaluation and long-term tracking module is used for generating an individual and group intervention effect report. By utilizing the system and the method, accurate, personalized, multi-dimensional and long-term intervention on the autism children can be realized.

The invention discloses a method and a system for training a dialogue intervention large model for children with autism, and the system comprises a data processing module which is used for converting original dialogue data from theme dialogue intervention records of children with autism and doctors into a high-quality multi-round dialogue data set; the model training module is used for finely adjusting an existing large language model on the basis of the high-quality multi-round dialogue data set so as to learn dialogue styles and intervention strategies in the high-quality multi-round dialogue data set, and an intervention large model which can follow an ABA application behavior analysis principle to perform theme dialogues with autistic children is obtained; and the model evaluation module is used for comprehensively evaluating the intervention effect of the intervention large model from the sentence dimension and the dialogue dimension so as to feed back the model fine adjustment process. According to the method, the large model following the ABA application behavior analysis principle can be trained to perform dialogue intervention with the autism children, the intervention effect is effectively improved, and accessible, efficient and low-cost social communication rehabilitation support is provided for the autism children.

The application discloses a method, device and diagnosis and treatment machine auxiliary system for autism child behavior recognition and motion prediction, the method comprises the following steps: (1) acquiring original skeleton video data of the child from different angles; (2) inputting the acquired original skeleton video data into a view adaptive conversion unit to adaptively convert to a consistent coordinate system to obtain the optimal representation of the human skeleton; (3) inputting the optimal representation of the human skeleton into a multi-scale feature extraction unit to extract high-performance spatiotemporal features of the human skeleton; (4) inputting the high-performance spatiotemporal features into a multi-task learning unit to perform behavior recognition and motion prediction on the human body, and obtaining behavior classification results and motion prediction results. The application can predict and intervene the future motion while recognizing the behavior.

The application discloses an interactive autistic child training system, method, electronic device and readable storage medium. The interactive autistic child training system comprises a training configuration module, a TOF radar, a multi-modal sensor fusion camera group, a gesture exercise recognition module, an entity interaction behavior recognition module and a prompt module. The training configuration module is used for configuring training actions / training elements corresponding to autistic life self-care skills. The TOF radar is used for collecting gesture exercise data of palms and knuckles. The gesture exercise recognition module is used for judging the gesture exercise data. The multi-modal sensor fusion camera group is used for collecting entity interaction behavior data of a trainee. The entity interaction behavior recognition module is used for judging the entity interaction behavior data. The prompt module is used for outputting an object according to a personalized ability model of the trainee. The application is used to improve the scientificity and systematicness of training.

The application relates to a medical engineering technology, in particular to a control method and system of a diagnosis and treatment robot for autistic children. The control method comprises the following steps: analyzing a human face video by using a deep learning model; generating a next-step intervention action recommendation list by using an LSTM state prediction model and a CPT action decision model; selecting or modifying one of the recommendation items and executing; retraining and updating the LSTM state prediction model and the CPT action decision model by using new interaction data. The control system comprises an execution module, a perception module, a memory module, an evolution module and a decision module. By introducing the quantification calculation of the concentration of the children and the valence and arousal of the facial expression, and by using the LSTM to predict the future state, the strategy can be adjusted in advance when it is predicted that the children will lose interest, so that the children's boredom or escape behavior is not easily caused, and meanwhile, the model is retrained and updated, so that the self-learning and adaptation capability is high, and the auxiliary diagnosis and treatment capability is relatively high.

The invention discloses an early risk prediction method for autistic children based on postures and facial expressions, and aims to construct a deep neural network framework fusing a facial expression image sequence and body posture motion information for the pain point that existing ASD early recognition depends on a subjective doctor observation scale and lacks objective quantitative indexes. A specific convolutional neural network (CNN) branch is designed to extract facial expression features, a time sequence modeling network (such as Transform) is utilized to analyze attitude sequence features, an attention mechanism is combined to fuse multi-modal information, and finally automatic and objective recognition of natural flowing fine abnormal expressions and engraving attitudes in social interaction of young children is realized. And an effective intelligent tool is provided for early screening and auxiliary diagnosis of ASD.

The invention discloses a social training system and method for children with autism based on multi-mode music interaction, and relates to the technical field of social training for children with autism. Visual, audio and behavior data are cooperatively integrated through the multi-modal data acquisition unit, facial micro-expressions, voice and intonation changes and body movement details of children in social interaction are comprehensively covered, and subtle emotion signals or behavior clues possibly missed by a single sensor are effectively avoided; multi-source information is weighted and integrated in real time by using an attention mechanism of a feature fusion module, a unified high-dimensional representation vector is constructed, and the recognition precision of social intent and emotional state is improved; music parameters are dynamically adjusted based on an analysis result, and children are guided to naturally practice target behaviors in the simulated social scene; music changes and story branches are seamlessly linked through the plot driving module, social rule embedding is enhanced, and it is ensured that the children gradually internalize social skills in the immersive environment.

The invention provides a rehabilitation training system for autistic children based on a virtual reality technology, and belongs to the technical field of medical rehabilitation, and the system comprises a doctor terminal which operates on a tablet device and is configured to respond to a doctor touch action to generate a scene generation instruction or a scene switching instruction; the patient end runs on the VR head-mounted display device and is configured to respond to the scene generation instruction and generate and display a mixed reality scene or a virtual reality scene; the patient end is further configured to respond to the scene switching instruction and adopt a panoramic scene gradual change technology for switching during scene switching; and the management end runs on a server or a webpage platform and is used for storing and checking the training data from the doctor end and the patient end. According to the method, switching of different scenes is linked through a panoramic scene gradual change technology, sensory stimulation and anxiety mood caused by instantaneous jumping of a picture to a child patient are avoided, the diversity of the scenes can be improved, the application scene adaptation degree in the rehabilitation training process is greatly improved, and the training safety and compliance are remarkably improved.

The application discloses an autism child social interaction behavior video image analysis method and system, belongs to the technical field of computer vision and medical image analysis, and comprises the following steps: a video acquisition step, wherein video images of a child and an assessor performing a structured social interaction task are acquired through a multi-angle camera; a human body posture estimation step, wherein two-person skeleton key point detection is performed and posture parameter sequences are calculated; a social signal detection step, wherein key social behavior events such as eye contact, social smile, directional gesture and body orientation response are identified; an interaction synchronism analysis step, wherein the time coupling degree of posture changes of both parties is calculated through cross-correlation analysis; and an evaluation report generation step, wherein a social interaction ability evaluation report is generated in the form of a radar chart by comprehensively considering multi-dimensional features. The application realizes objective quantitative evaluation of the social interaction ability of an autism child and improves the consistency and repeatability of evaluation.

The application discloses an autism children attention prediction method based on a Mamba-Unet structure, and comprises the following steps: firstly, a multi-modal data set for autism children attention prediction is constructed; secondly, an autism children attention prediction model based on the Mamba-Unet structure is constructed and trained; and finally, a non-typical visual attention prediction is performed on a to-be-detected image by using the trained model to obtain a prediction result. The model comprises a pretreatment module, an encoder part, a bottleneck part, a decoder part and an output part which are connected in sequence, wherein an efficient adaptive visual state space block, a frequency domain attention module, a conditional visual state space block and a conditional attention fusion module are arranged; and the output of the encoder part is further transmitted to the decoder part through a jump connection. The method can significantly improve the accuracy of autism spectrum disorder non-typical visual attention prediction.

The application discloses a system and method for autism risk assessment and early warning, and relates to the field of computer and electroencephalogram signal processing. The system comprises an electroencephalogram data acquisition module, a feature extraction module, a risk assessment module and a risk early warning module. The preset task state covers resting state, local motion state and social motion state. In the resting state, the user observes low-speed target objects in the display picture for a preset time length. In the local motion state, the user observes the first type of target person completing local limb motion. In the social motion state, the user observes the second type of target person waving hands to say hello. By collecting electroencephalogram data in each task state and extracting features, the autism risk probability and the corresponding risk level are determined, and the early warning is triggered according to the preset logic rules. The system does not need frequent language interaction with children, is suitable for young children, significantly reduces the screening manpower and time cost, and is convenient for popularization in areas with scarce medical resources.

The invention discloses a domain adaptive space-time electroencephalogram modeling emotion recognition method and system for autism. The method comprises the steps that typical development child electroencephalogram signal fragments with labels serve as source domain data, autism child electroencephalogram signal fragments without labels serve as target domain data, and the source domain data and the target domain data serve as training data together; training data are input into the emotion recognition model for training optimization, a graph is dynamically constructed on the basis of physical coordinates of a channel, a graph structure is dynamically updated on the basis of the graph, and spatial topological structure features of the data are learned through graph convolution; meanwhile, time sequence features are obtained through a time sequence feature extraction module, and then time-space features are fused through a multi-head attention mechanism; and after fusion features are obtained, domain invariant features of electroencephalogram fragments are fully learned in a feature space through a transfer learning method. According to the method, the emotion recognition task of the autism children can be completed only by using the labels of the typical development children.

The interesting rehabilitation training device comprises a bottom plate, vertical plates are fixedly mounted on the left side and the right side of the top of the bottom plate correspondingly, connecting frames are fixedly mounted on the left side and the right side of the top of the bottom plate and located on the inner sides of the vertical plates correspondingly, and a turnover mechanism is arranged at the top of the bottom plate; the turnover mechanism comprises a double-shaft motor, an output shaft, a transmission wheel, a transmission belt, a driven wheel and a sliding sleeve, through the arrangement of the turnover mechanism, richer and more personalized training experience can be provided for autistic children, the children are helped to make contact with various different toy teaching aids on one device, the game experience of the children is enriched, and the interestingness of the children is improved. And secondly, the all-in-one design can promote the sensory and cognitive development of the children, so that the children can understand and accept different types of stimulation more easily, and the development of social skills of the children is also facilitated because the children can be used together with other children to play interactive games.

The invention discloses an autistic child evaluation system and device based on an electroencephalogram large model, the system comprises an electroencephalogram collection module, a preprocessing module, a model training module and a control module, the electroencephalogram collection module is used for collecting a first electroencephalogram data set of a non-autistic child group and a second electroencephalogram data set of an autistic child group; the preprocessing module is used for preprocessing the first electroencephalogram data set and the second electroencephalogram data set to obtain a third electroencephalogram data set and a fourth electroencephalogram data set; the model training module is used for determining a target pedestal model according to the third electroencephalogram data set, the fourth electroencephalogram data set and a preset pedestal model; the electroencephalogram acquisition module is also used for acquiring to-be-evaluated electroencephalogram data corresponding to the target child; and the control module is used for determining a target evaluation result corresponding to the to-be-evaluated electroencephalogram data according to the target base model. By adopting the embodiment of the invention, the reliability of the autism evaluation result can be improved.

The application discloses an autism child stereotyped repetitive behavior image recognition method and system, and belongs to the technical field of artificial intelligence and image recognition. The method comprises the following steps: acquiring whole body movement video image data of an autism child, and extracting a movement trajectory sequence through a human body skeleton key point detection network; detecting a periodic repetition mode in the movement trajectory through a time sequence analysis algorithm, and recognizing five stereotyped behavior types, i.e., hand slapping, body shaking, rotating behavior, arranging articles and repeatedly opening and closing a door; establishing a stereotyped behavior feature library, and extracting quantitative parameters of a movement period, a movement amplitude, a continuous time length and a triggering situation; calculating a stereotyped behavior comprehensive score, supporting multi-period comparison and analysis to evaluate intervention effects; and generating an analysis report to assist clinical diagnosis and rehabilitation training scheme formulation.

This invention discloses a method for screening autism children based on eye movement detection using ordinary video. The method comprises a video acquisition unit, a preprocessing unit, an eye movement estimation unit, a scene analysis unit, a feature quantization unit, a screening inference unit, and a display unit. It uses an ordinary camera to capture facial video of children, integrates head posture compensation to achieve label-free eye movement tracking, automatically segments dynamic social regions and extracts autism-specific eye movement features, completes risk screening through a lightweight model, and outputs a visual report. This invention requires no professional eye tracker, no head fixation, is suitable for children with low cooperation, and is fully automated. It has the advantages of low cost, strong robustness, good interpretability, and ease of implementation at the grassroots level, making it suitable for large-scale early screening of autism children.

The invention relates to the technical field of autism child behavior analysis, and provides an autism child daily behavior data acquisition method and system, and the method comprises the steps: collecting the three-axis acceleration and three-axis angular velocity of an inertial measurement unit in real time, and taking the three-axis acceleration and three-axis angular velocity as original data; extracting feature data of the original data; according to the feature data, judging the physical consistency of the original data, and distinguishing a real limb movement signal and a sensor relative movement signal; decomposing the original data into a first information stream and a second information stream; based on the first information flow, identifying a behavior pattern of limb movement of the autism child; extracting features of the second information flow to obtain second information features; the second information feature is used for representing the intensity of the relative motion of the sensor corresponding to the relative motion signal of the sensor; calculating a behavior intensity index according to the second information feature; and outputting the behavior pattern and the behavior intensity index. The method has the effect of improving the accuracy of collecting the daily behavior data of the autism children.

The invention is suitable for the technical field of behavior recognition, and provides a deep learning algorithm-based autism child carving behavior recognition method and system, and the method comprises the steps: collecting multi-view child carving behavior data, and carrying out the standardization processing; selecting representative carving behaviors in the multi-view data set to obtain video clips; the data set is processed based on a frequency-Colmogov-Arnod network, and frequency characteristics of the depicted behavior children are extracted; generating a feature map through algebraic normalization layer compression coding; relative entropy is introduced into cross entropy loss to serve as a regular term, the difference between output probability distribution and prior distribution is restrained, and different types of gradients are distinguished through asymmetry of the relative entropy. According to the method, frequency information is introduced into the KAN network, frequency components in autism behaviors are extracted, and the prediction precision is improved; an algebraic normalization layer is used to relieve the condition of complex calculation in an autism scene; the method solves the problem that the inscription behaviors in the autism scene are many and miscellaneous and are not easy to identify.

The invention relates to an AI interaction system and interaction method for nursing and social contact of autistic children, the system comprises a robot terminal, a cloud AI platform and a parent / therapist terminal which are connected in sequence through a communication module, and the robot terminal is used for collecting multi-modal data of the autistic children and sending the multi-modal data to the cloud AI platform. The data are uploaded to a cloud AI platform after being subjected to multi-modal data processing; the parent / therapist terminal is used for receiving the interactive abstract and the key data and configuring personalized parameters to carry out personalized adjustment and optimization on an AI model in the cloud AI platform; and the cloud AI platform is used for analyzing the processed multi-modal data by adopting the personalized and optimized AI model, generating an interaction abstract and key data, generating an interaction script, transmitting the interaction abstract and the key data to the parent / therapist terminal, transmitting the interaction script to the robot terminal, converting the interaction script into voice through the robot terminal, and outputting the voice. Compared with the prior art, the method has the advantages of individuation, interaction diversity and the like.

The invention discloses a multifunctional auxiliary chair for venipuncture of autistic children, relates to the field of nursing instruments for the department of neurology of children, and aims to solve the problems that existing equipment is not adaptive to autistic children, and the autistic children are sensitive in feeling, weak in emotion regulation and split in function. The auxiliary chair comprises a chair body, a limb restraining assembly and an animation guiding module, and is matched with scene adaptation structures such as height adjustment and lower limb folding. The limb restraining assembly adopts a flexible wrapping layer with a pressure sensor and an adjustable restraining belt, and safe flexible restraining and pressure real-time monitoring are achieved; the animation guiding module is electrically connected with the pressure sensor, the puncture process, pacifying and science popularization animations can be dynamically switched through the display screen, character prompts can be popped up, the constraint strength can be synchronously adjusted, and emotion guiding and constraint linkage is achieved. The device effectively relieves the puncture conflict emotion of autism children, guarantees constraint safety, adapts to different heights, body types and puncture scenes, improves the one-time puncture success rate, reduces the operation risk and meets the clinical nursing requirements.

The application belongs to the technical field of autism children rehabilitation evaluation, and discloses an autism children rehabilitation evaluation method, system and device based on an MREP mode, which converts the performance of children on target evaluation projects into quantitative feature data through structured behavior quantitative coding, calculates the abnormality of children in four dimensions of medical treatment, rehabilitation, education and family based on the data, screens out a primary intervention target set in combination with development lag judgment, automatically matches a structured intervention scheme for each target and sorts and executes the structured intervention scheme by querying a preset evidence-based intervention mapping matrix, collects process data of institutions and families within an intervention period, and calculates a progress index of each target according to the process data, and obtains a comprehensive evaluation value by weighting the average mastery degree of each dimension target and the overall target achievement rate. The application realizes the full-process data of collection, intervention and effect evaluation, and improves the accuracy, scientificity and efficiency of rehabilitation services.

The application belongs to the field of autism children emotion regulation, and provides a method and system for autism children emotion recognition and regulation based on wearable devices, and the process is as follows: after the physiological data and facial expressions of autism children are input into a mixed multi-layer neural network after data processing, emotion recognition results are obtained; if negative emotions are recognized, the inflatable vest performs inflation action to relieve the bad mood of the children. The application uses multi-modal data to identify the emotions of autistic children, which is conducive to fully mining the real emotional characteristics of the children and improving the emotion recognition accuracy. At the same time, the application uses the sensitivity of physiological signals to emotional fluctuations to realize the functions of emotion warning and early intervention. In addition, the application systematically integrates various wearable devices, which can continuously, intelligently and effectively regulate the emotions of autism children.

The embodiment of the application discloses a multi-modal emotional interaction system and method, the system comprises a client and a server, through the cooperation of the client and the server, the application behavior analysis (ABA) theory is used as a basic framework, the behavior of the child is understood and intervened by the parents, and through the simulation scene practice, the parents are guided to understand the details of the intervention behavior, a structured roadmap can be provided for the parents, the behavior of the child is helped to be understood and managed, and meanwhile, the theoretical concept is combined with the practical application, the ability of the parents to provide customized intervention for the needs of the child is improved. Can be used for solving the problems that the information source of the current autism child care is unreliable, personalized content is insufficient, and there is a gap between theoretical knowledge and actual operation.

The application is suitable for the technical field of behavior recognition, and provides an autism child stereotyped behavior recognition classification system and method based on RMG-transformer, which comprises the following steps: video data acquisition and preprocessing; repeated action segment detection; super image construction; feature extraction and behavior classification; and recognition result output. The repeated action detection submodule and the super image submodule are innovatively introduced in the original Transformer network architecture, and the loss function is further optimized, so that the efficiency can be improved according to the characteristics of the stereotyped behavior repetition, and the time sequence characteristics are enhanced, and the recognition ability of continuous actions is improved.

The invention discloses an autism child rehabilitation intervention system and method based on an AI face changing technology, and the system comprises a data collection module which is used for collecting the facial features and limb motions of a child patient; the AI face changing module is used for replacing the face of a standard virtual character with the face of the child patient so as to generate a personalized virtual character of the self image of the child patient; the action library module is used for storing standardized actions and corresponding emotion expression data; and the interactive feedback module is used for monitoring and comparing the action of the child patient with the target action in real time and providing visual or auditory feedback. The intervention method comprises the steps of collecting a face image of a child patient and generating a virtual image; playing the target action demonstrated by the virtual image; actually measuring the simulation action of the child patient and calculating the simulation proportion and the simulation accuracy; and dynamically adjusting the next difficulty or reward feedback according to the result. The method has high compliance, the acceptance of ASD children to the virtual character of the self image is obviously higher than that of a traditional method, and the method is efficient and accurate.

The invention provides an attention network-based autism children video screening method and system in order to solve the problems of difficulty in early recognition and insufficient capture of subtle behaviors in the existing autism screening of children, and the screening method comprises the following steps: designing a cross-channel sparse attention mechanism, and applying sparse constraints among different modes to obtain a video screening result of the autism children. The attention between the visual channel and the semantic Token only keeps key relation, redundant interference is reduced, and the sensitivity to an abnormal social behavior mode is improved; then, a multi-granularity space-time attention encoder is constructed, the multi-granularity space-time attention encoder comprises local behavior attention, global behavior attention and behavior prior guide attention and is used for modeling short-term action changes, long-term behavior patterns and autism related behavior prior, and a joint enhanced semantic feature Finc (t) is output; and finally, in the step 4, an adaptive fusion gating unit is introduced, visual features and enhanced semantic features are dynamically integrated, a final feature representation Fcls (t) is formed, Bayesian uncertainty estimation is matched, and the screening reliability is improved.

This invention discloses a method for guiding picture book reading for children with autism based on emotional segmentation and visual silent cues. The method includes: extracting and fusing emotional features from the picture book text and illustrations to construct an emotional arc and segmentation boundaries; collecting children's responses to establish a perceptual sensitivity vector; calculating the cumulative load value of subsequent pages based on an emotional aftershock model when turning pages, generating a risk timeline; collecting emotional observation values ​​in real time, using Kalman filtering to track the baseline and calculate the deviation value; combining the deviation value and the risk timeline to determine trigger rules, generating cue parameters based on the intervention urgency, and executing visual silent cues; after the session, bidirectionally correcting the emotional labeling values, and using the least squares method to fit the aftershock sequence and update the individual attenuation coefficient. This invention achieves preventative intervention through the cumulative effect of emotion, and the bidirectional self-correction closed loop improves accuracy, making it suitable for assistive therapy through picture book reading for children with autism.

The invention discloses a risk assessment and early warning system and method for autistic children, and relates to the field of computers and electroencephalogram signal processing. The system comprises an electroencephalogram data acquisition module, a feature extraction module, a risk assessment module and a risk early warning module. A preset task state covers a resting state, a local motion state and a social motion state. In the resting state, a user observes a low-speed target object in a display picture and continues for a preset duration. Observing that the first type of target person completes local limb movement in a local movement state; and in the social motion state, observing the picture of the second type of target person moving upper limb greeting. The autism risk probability and the corresponding risk level are determined by collecting all task state electroencephalogram data and extracting features, and then early warning is triggered according to a preset logic rule. The system does not need frequent language interaction with children, adapts to young child groups, remarkably reduces the screening manpower and time cost, and is convenient to popularize in regions deficient in medical resources.

The invention discloses an autistic child emotion recognition method fusing a knowledge graph and a physiological signal, and the method comprises the steps: cleaning data, recognizing an entity from the cleaned medical record data to obtain medical record information, searching a neighbor node in a mental disorder knowledge graph through employing the medical record information, constructing a sub-graph most related to the medical record information based on the neighbor node, and carrying out the recognition of the emotion of an autistic child. Performing graph embedding processing on the sub-graphs to obtain a low-dimensional embedding vector set; constructing a vertex set based on the low-dimensional embedded vector and the physiological signal features, constructing hyperedges based on vertexes, constructing a multi-modal incidence matrix of a hypergraph by using the vertexes and the hyperedges, and constructing a multi-modal hypergraph by using the multi-modal incidence matrix; updating the multi-modal hypergraph and optimizing and updating the target function; obtaining a predicted emotion label based on the target function and carrying out online learning; the limitation that a single-mode signal cannot comprehensively reflect the emotional state is overcome; the problem of modal heterogeneity in multi-modal signal fusion is solved.