45results about How to "Addressing Individual Differences" patented technology

The present invention provides a dynamic model directed angiography three-dimensional reconstruction method, pertaining to an intersectional field of digital image processing and medicine imaging, for satisfying special requests of auxiliary detection and surgery guidance for cardiovascular disease in clinical medicine. The invention includes steps of angiography image preprocessing, vascellum segmentation, vascellum skeleton and radii extraction, model directed vascellum base element recognition, vascellum matching and vascellum three-dimensional reconstruction. The invention also provides a cardiovascular dynamic model construction method including steps of cardiovascular slice data extraction, cordis static and dynamic model building, and static and dynamic model building. According to the invention, good angiography three-dimensional reconstruction result may be obtained, which effectively assists detection and surgery guidance for cardiovascular disease, thereby satisfying clinical requests.

The invention discloses a gesture recognition method based on an electromyographic topographic map. The method comprises the steps of (1) data collection, wherein upper arm muscle surface electromyographic signals of different gestures are collected through array surface electromyographic electrodes; (2) data preprocessing, wherein the collected surface electromyographic signals are preprocessed; (3) generation of the electromyographic topographic map; and (4) deep convolutional neural network model training and gesture recognition for generation of a feature image of the electromyographic topographic map, wherein the electromyographic topographic map is converted into a 64*64 grayscale image first, and then ZCA whitening preprocessing is used to generate the feature image; a corresponding convolutional neural network model structure is designed according to characteristics of the electromyographic topographic map, and a model is constructed; and test set data is input into a trained network model for gesture recognition classification. Through the method, the same gesture can be made to different subjects to generate similar electromyographic topographic maps, and therefore the problem of individual difference of surface electromyographic signals is effectively solved.

The invention relates to a facial feature based driver attention state detection method. The detection method is characterized by comprising the following steps of performing PERCLOS based driver fatigue state detection; performing blink frequency based driver fatigue state detection; performing sight line distribution based driver attention distraction state detection; and sending out early warning for driver fatigue and attention distraction state, wherein the early warning for fatigue or attention distraction state is carried out continuously for a certain time to avoid frequent switching between different states; the early warning priority is as follows: fatigue driving is greater than attention distraction that is greater than normal driving; and the attention state of the driver can be accurately detected through the physiological reaction characteristics of the driver, and early warning is sent out for driver fatigue and the attention distraction state; collection and extraction of the facial features of the driver are not included; the sampling efficiency of the adopted driver facial feature collection and extracting system is fHz; and the frequency for reading the driver facial feature is f<0> Hz.

The invention relates to a matching method for three-dimensional coronary angiography reconstruction, which belongs to the crossing field of digital image processing and medical imaging, and aims to meet the special requirements on the auxiliary detection and the surgical navigation of angiocardiopathy in clinical medicine. The method comprises two steps, i.e. model guidance contrastographic picture vessel segment matching and contrastographic picture puncta vasculosa matching. The matching method uses a multi-scale vascular tree model so as to better guide vascular matching, simultaneously can well solve the problem of individual difference, and improves the matching precision of vessel segments and puncta vasculosa through the method of iterative matching and largest declination of searching potential energy. The matching method for three-dimensional coronary angiography reconstruction can obtain very good angiography matching results, thereby solving the difficult problems in the automatic three-dimensional reconstruction by the multi-view angiography map.

The invention provides a mechanical device and a method for achieving constant-speed, constant-pressure and non-missing plain canning of a B ultrasound probe. The device comprises a guide rail assembly and a guide rail assembly bracket; the guide rail assembly comprises an X axial guide rail, a Y axial guide rail, a Z axial guide rail, and a rotatable B ultrasound probe clamp which is mounted at the bottom end of the Z axial guide rail; a pressure sensor and an angle measuring sensor are arranged on a lower bracket of the Z axial guide rail; the guide rail assembly is connected with the guide rail bracket through a rotating shaft and can be rotated around the rotating shaft; the guide rail assembly rotates to drive a motor and a rotating speed reduction device; the transmission mode of the guide rail is performed through a precision lead screw or a gear belt; the B ultrasound probe is driven by a driving mechanism to axially move at constant speed in the X, Y and Z directions; meanwhile, the probe is swung under the effect of the probe clamp rotating mechanism; a track assembly is driven by the rotating mechanism to rotate to adjust the vertical relationship between the ultrasound direction and the detection surface, so as to achieve constant-pressure, constant-speed, non-missing and overlapping-free scanning of the B ultrasound probe.

A control technology for a synchronous motor for suppressing rotational pulsation caused by variation in individuals without making a control algorithm complex is provided. In a motor drive system which is a control device for a synchronous motor, in order to suppress the pulsation component of N times as high as the AC frequency for driving the synchronous motor, a controller in which the phase property of the disturbance response of the controller with respect to the pulsation frequency is within ±45° is arranged. Therefore, the torque pulsation component generated from distortion in induction voltage or variation between phases is suppressed.

The invention discloses a three-dimensional fine blood vessel reconstruction method. The three-dimensional fine blood vessel reconstruction method comprises the following steps of 1, loading a three-dimensional fine blood vessel original image; 2, performing blood vessel property-based image preprocessing; 3, performing blood vessel property-based characteristic matching; and 4, performing image filling and boundary processing. The invention furthermore discloses a three-dimensional fine blood vessel reconstruction system. With the adoption of the three-dimensional fine blood vessel reconstruction method and system provided by the invention, a filling region for repairing a missing part is obtained by mapping a position relationship and calculating an overall matching degree, and the missing part is repaired through the filling region to obtain a complete high-precision image, so that the accuracy of three-dimensional fine blood vessel reconstruction is high.

An image scanner is provided with a super-resolution processing function for generating a high-resolution image using multiple pieces of low-resolution image data. The scanner device then measures an initial MTF value without performing the super-resolution processing, refers to an MTF properties table in which is registered a number of pieces of low-resolution image data for a scanner device of that type to achieve a pre-set target MTF, and registers the number of images that corresponds to the initial MTF value in the scanner, if that number is present, and registers an interpolated value if that number is not present. The image scanner then performs the super-resolution processing using the registered number of pieces of low-resolution image data, and outputs image data of a target resolution.

The invention aims at providing an anaesthetic balance control device. The device comprises a data filter module and a drug balance control module. Pre-stored electrocerebral indexes, noxious stimulation data drug usage data are divided into electrocerebral index-sedative data and noxious stimulation data-sedative data. Corresponding drug target control indexes are selected according to electrocerebral index and noxious stimulation data changes in certain period, threshold values are preset, and injection of muscle relaxants are increased or stopped based on the set threshold values. Meanwhile, bispectral index, noxious stimulation data, neuromuscular retarding depth data change tendency slope and the like can be analyzed dynamically, and transverse balance adjusting control is performed on three signals. The invention further provides a work method of the device. According to the device and the method, the problem the imbalanced usage of three drugs is solved, effects of individual differences are prevented, the safety during anesthesia is guaranteed to the largest extent, and good social values and application prospects are provided.

A control technology for a synchronous motor for suppressing rotational pulsation caused by variation in individuals without making a control algorithm complex is provided. In a motor drive system which is a control device for a synchronous motor, in order to suppress the pulsation component of N times as high as the AC frequency for driving the synchronous motor, a controller in which the phase property of the disturbance response of the controller with respect to the pulsation frequency is within ±45° is arranged. Therefore, the torque pulsation component generated from distortion in induction voltage or variation between phases is suppressed.

The invention discloses a cross-subject EEG cognitive state recognition method based on a prototype clustering domain adaptation algorithm. According to the method, the concept of category domains isintroduced, on one hand, on the basis of multi-source domain alignment of labels, feature distribution differences between different categories are considered, structural fine-grained alignment underthe category conditions between different source domains in a feature space is researched, and the problem of category imbalance in the multi-source domains is converted into a mode of the category domains; and prototype theoretical clustering alignment between the source domain and the target domain is carried out, i.e., clustering between similar source domains is carried out on the target domain by taking a dynamic adjustment prototype center as a constraint, and similar features and sparse heterogeneous features between the domains are realized, wherein the former realizes intra-domain class conditional structure feature alignment, and the latter realizes global fine-grained structure feature alignment. According to the invention, the method can be compatible with category balance andimbalance, effectively solves the problem of individual difference of electroencephalogram signals in the field of brain cognitive calculation, has high generalization ability, and can be well suitable for clinical diagnosis and practical application.

The invention relates to a brain-computer interface system based on auditory attention and multifocal electrophysiology and an implementation method. The brain-computer interface system based on auditory attention and multifocal electrophysiology is characterized in that the brain-computer interface system comprises an auditory stimulator module, a signal acquisition module, a signal processing module and an execution module; the auditory stimulator module can generate sound sources with multiple frequencies and multiple sound tracks simultaneously, and the sound sources are distributed at different positions of virtual auditory space; all the sound sources are controlled by a same M sequence, and the sound sources produce sounds when the M sequence is 1 and do not produce sounds when theM sequence is 0; the signal acquisition module collects and acquires mixed electrophysiological signals of auditory response through sensors, and signal amplification and simulation / digital conversionare conducted; the signal processing module conducts quick mutual closing calculation on the mixed electrophysiological signals and the M sequence, and auditory response signals corresponding to allthe sound sources are acquired; and then characteristic value extraction is conducted on the auditory response signals of all the sound sources, characteristic vectors are classified through a distinguishing and classifying module, and the sound sources drawing the attention of a tester are decoded, so that the mind of the tester is determined, and the execution module is controlled to execute corresponding action.

The invention relates to the technical field of intention recognition, and provides an upper limb wearable transfer robot motion recognition system based on multi-signal fusion. The upper limb wearable transfer robot motion recognition system comprises a sensor system and a data processing system. The sensor system comprises a surface dry electrode myoelectricity sensor, a six-axis inertial sensorand a silica gel air bag connected with an air pressure sensor. The data processing system is an upper computer integrating a human body physiological information calculation module, an electromyographic signal characteristic value calculation module, an upper limb joint angle calculation module, a four-stage processing module and an impedance regulator. The four-stage processing module comprisesa neural network module, a principal component analysis module, a moving average filtering module and an If-Then decision maker module. The output end of the sensor system is electrically connected with the input end of the data processing system, and the output end of the data processing system is electrically connected with the input end of a robot controller. According to the invention, the individual difference problem of physiological signals can be solved, and the accuracy of intention recognition and the effectiveness and safety of man-machine interaction are improved.

The invention discloses a multi-data-source data acquisition method. The method comprises the steps of obtaining to-be-acquired data through a preset protocol; when the to-be-collected data is the data of the identification code device, scanning an identification code on the identification code device through a scanning device to obtain the to-be-collected data; when the to-be-collected data is inthe running state of the industrial equipment, obtaining the to-be-collected data through a sensor; when the to-be-collected data is the questionnaire data, obtaining the to-be-collected data througha mobile terminal; processing the to-be-collected data, and displaying the to-be-collected data through a display screen; and uploading the to-be-collected data to a cloud, so that the cloud sends the to-be-collected data to an upper computer. The invention also discloses a multi-data-source data acquisition device and a multi-data-source data acquisition system. By adopting the embodiment of theinvention, the data of a plurality of data sources can be automatically acquired, and the instantaneity, reliability and safety of equipment communication are ensured.

The invention provides an online gesture recognition method for the myoelectricity individual-difference problem. The method comprises the following steps: (1) establishing a gesture myoelectricity data set to obtain initial data; (2) training an initial classification model; (3) carrying out KKT (Karush-Kuhn-Tucker) discrimination of newly added samples; (4) using DBSCAN density clustering to obtain a preliminary training set; and (5) carrying out secondary screening to obtain a final training set. The online gesture recognition method for the myoelectricity individual-difference problem provided by the invention combines a DBSCAN density clustering algorithm, improves an original KKT-SVM incremental-learning method, and is used for online myoelectricity gesture recognition to solve the individual-difference problem.

The invention provides a decorated metal visual inspection system, which comprises a feeding mechanism, a visual acquisition and inspection mechanism, a waste rejecting mechanism and a receiving mechanism, wherein the feeding mechanism, the waste rejecting mechanism and the receiving mechanism are sequentially connected by means of a conveying belt; the waste rejecting mechanism is electrically connected with the visual acquisition and inspection mechanism; the visual acquisition and inspection mechanism comprises a visual acquisition module and a visual inspection mechanism which are electrically connected with each other; the visual acquisition module is arranged above the conveying belt positioned between the feeding mechanism and the waste rejecting mechanism; and the waste rejecting mechanism is used for carrying out waste rejection processing according to a corresponding electrical signal which is transmitted from the visual inspection module through determining defective decorated metal by detecting decorated metal patterns acquired by means of the visual acquisition module. The decorated metal visual inspection system solves the technical problem of low qualified rate sincedefective products cannot be detected and flow to the customers due to visual fatigue and individual difference of long-term manual operation. The invention further provides a decorated metal patterninspection method.

The invention relates to a similar case recommendation method based on comprehensive feature representation and an improved width-depth model, and belongs to the technical field of computer natural language processing. According to the invention, the method comprises the steps: firstly, obtaining comprehensive features described by medical records through a comprehensive feature representation model, and performing disease screening; secondly, processing the discrete features by adopting a cross feature method, inputting the discrete features into a linear part, combining comprehensive features described by medical records with shallow model features, and inputting the combined features into a recommendation sorting part taking a gate loop unit as a core; and finally, on the basis of hundreds of candidate cases, outputting dozens of case recommendation items. According to the method, the personalized case recommendation is realized, a recommendation sorting algorithm model combining atraditional shallow linear model and a deep network model is provided, and the accuracy of similar case recommendation is improved.

The invention adopts a method of cryptocaryon irritans locally-infected epinephelus coioides one-sided skin for immunity for the first time, adopts an optimized infection concentration and infection duration time to construct an immune model of the cryptocaryon irritans locally-infected epinephelus coioides one-sided skin for the first time, thoroughly eliminates an adverse effect brought by ''congenital differences'' in the process of local skin mucosal immune response in fish, and provides an ideal immune model for researching a local mucosal immune response law. The construction method forthe immune model mainly comprises the following five steps: (1) collecting live larvae of the cryptocaryon irritans and calculating a concentration; (2) grouping the epinephelus coioides and performing domestication; and (3) performing local skin immunization; (4) measuring infection intensity; and (5) performing treatment on locally-infected fish bodies.

The invention discloses a method for immunization by using cryptocaryon irritans to locally infect the unilateral gill of epinephelus coioides for the first time, the optimized infection concentrationand infection duration are used to construct an immune model of the cryptocaryon irritans to locally infect gill tissues of the epinephelus coioides for the first time, the negative effects caused bycongenital difference in the immune response process of fish local gill tissue mucous membranes are completely eliminated, and an ideal immune model is provided for studying the local mucous membraneimmune response rule. A construction method for the immune model mainly comprises five steps of (1), collecting cryptocaryon irritans live larvae, and conducting concentration calculation; (2), conducting grouping and domestication on the epinephelus coioides; (3), conducting local gill tissue immunization; (4), measuring infection intensity; (5) treating a fish body after local infection.

The invention discloses an adversarial partial domain adaptive cross-subject EEG emotion recognition method based on clustering, and the method comprises the steps: calculating a class cluster center through employing the features of a source domain sample, taking a real tag of a source domain as a class cluster tag, introducing a consistency matching algorithm and a cross-domain clustering consensus index, and carrying out the recognition of the consensus partial domain adaptive cross-subject EEG emotion. Using Kmeans clustering to obtain a class cluster label and a class cluster center corresponding to a label-free target domain sample, carrying out consistency matching on a source domain class cluster center and a target domain class cluster center, for two successfully matched class clusters, distributing the source domain label to the target domain class cluster with the same semantic meaning, and carrying out the matching of the source domain class cluster and the target domain class cluster with the same semantic meaning; meanwhile, cross-domain clustering consensus indexes are calculated to achieve search of the optimal number of target domain class clusters, correlation of common classes and separation of private classes of the source domain and the target domain are finally achieved, the feature space distribution structure of unlabeled data is fully considered, high universality is achieved, the model training efficiency can be greatly improved, and the method is suitable for large-scale popularization and application. And technical support is provided for clinical application.

An oral cavity palate soft and hard tissue segmentation method based on an attention mechanism and integrated registration comprises the steps that firstly, CBCT images are acquired, and after data labeling, the images are divided into a training set, a verification set and a test set; secondly, inputting the training set into a built oral palate soft and hard tissue segmentation network model; in addition, a random augmentation method is added during network training, and random scale and different transformation type enhancement is carried out on input data; predicting a tissue segmentation result, and quantitatively evaluating a prediction effect of the model on a test set; and finally, performing integrated registration according to a multi-palate soft and hard tissue segmentation result. The oral cavity palate soft and hard tissue segmentation and registration method fills the blank of oral cavity palate soft and hard tissue segmentation and registration, solves the problem that tissue segmentation is not accurate enough, shortens the time for searching implantation sites of different cases, and provides technical support for case analysis and design of orthodontic implantation nails.

The present invention provides a recommendation system scoring prediction method based on a cloud model facing the social network, belonging to the data excavation and information retrieval field. The method employs a data set in a scoring social network site to collect a user historical scoring record and the user social relation. Aiming at the user scoring subjectivity, an inverse cloud generator constructs scoring cloud for the user scoring, a comprehensive cloud model is employed to fuse all the user scoring cloud to generate a father cloud, a new score is generated under the father cloud, and the score is used for the clustering of the users to discover the user similar colony. A multi-rule comprehensive prediction method is constructed based on the Gaussian transformation through introducing a membership and high-dimensional cloud model and combination of the user social relation to overcome the scoring data sparse problem.

The invention provides an artificial intelligence processing method supporting online learning and a processor. By adopting a double-judgment mode combining adaptive template matching, adaptive high / low threshold judgment and neural network judgment, the system has an online learning capability, can adapt to time sequence signal characteristics of using equipment and users, solves the problem thatthe common time sequence signal characteristics have individual differences, and effectively improves the automatic classification accuracy. Compared with a traditional neural network back propagation algorithm for achieving hardware implementation of online learning, the method is small in parameter quantity and operand, only a small amount of data is stored, and compared with online learning ofthe neural network, the power consumption and the needed storage quantity are greatly reduced. All parameters in the online learning hardware module can be freely configured according to different application requirements, so that an online learning algorithm can be flexibly deployed on a processor, and the online learning hardware module is suitable for application scenes of automatic classification of various time sequence signals.

The invention discloses a constant-power aging circuit for an output end of a photoelectric coupler, and relates to the technical field of component aging. The constant-power aging circuit comprises a power supply V1, a power supply V2, an input end constant current module, an optocoupler to be aged, an output end current-limiting voltage difference-stabilizing module, an output end constant current control module and an output current sampling module. One end of the input end constant current module is connected with the positive electrode of the V1, the other end of the input end constant current module is connected with the input unit positive electrode end of the optocoupler to be aged and the positive output end of the output end constant current control module, and one end of the output end current-limiting voltage difference-stabilizing module is connected with the positive electrode of the V2, and the other end of the output end constant current module is connected with the positive electrode end of the output unit of the optocoupler to be aged. One end of the output current sampling module is connected with the negative electrode end of the output unit of the optocoupler to be aged, and the other end is grounded; the input end of the output end constant current control module is connected in parallel with the two ends of the output current sampling module, and the output end is connected in parallel with the two ends of the input unit of the optocoupler to be aged. The problems that an existing aging circuit is unstable in aging power, uncontrollable in quality, high in cost and the like are solved, and the circuit is widely applied to aging circuits of photoelectric couplers.

The invention relates to the field of biological products, in particular to a cell collection shaking table and a cell collection method. The invention provides the cell collection shaking table, which comprises a box body, a driving device and a tray assembly, wherein a closed collection area is arranged in the box body, the tray assembly is arranged in the collection area, a cell factory is arranged on the tray assembly, the driving device drives the tray assembly to rotate circumferentially with a set radius; the invention also provides a cell collection method adopting the cell collectionshaking table, wherein after the cells are cultured in the cell factory, the cell factory is fixed on the tray assembly, the tray assembly rotates for a set time with a set radius and a set rotating speed, and the set rotating speed is 100-300r / min. The cell collection shaking table and the cell collection method are suitable for cell collection in a cell factory and have simple structure, simpleand convenient operation, high collection efficiency and good collection effect.

The invention discloses a method and device for diagnosing equipment status. The method includes: processing the acquired historical samples to obtain typical sample fault classifications and their characteristic values; performing similarity judgment on the tested samples to obtain preliminary fault classifications; Obtaining the sample feature value corresponding to the preliminary classification of the fault, performing weighted calculation to obtain a similarity parameter, and determining the fault classification of the tested sample; when the fault classification is a single type, determining the fault classification as the The fault diagnosis classification of the tested sample; if it is not a single type, create an accurate model, input the preliminary fault classification into the accurate model for judgment, and obtain the fault diagnosis classification of the tested sample. The invention achieves the purpose of improving the accuracy and precision of equipment state diagnosis.

The three-dimensional reconstruction method of angiography guided by dynamic model belongs to the intersection field of digital image processing and medical imaging, and its purpose is to meet the special requirements of auxiliary detection of cardiovascular diseases and surgical navigation in clinical medicine in my country. The invention includes the steps of contrast image preprocessing, blood vessel segmentation, blood vessel skeleton and radius extraction, model-guided blood vessel primitive identification, blood vessel matching and blood vessel three-dimensional reconstruction steps. The present invention also provides a method for establishing a dynamic cardiovascular model, which includes the steps of extracting cardiovascular slice data, the steps of static and dynamic modeling of the heart, and the steps of static and dynamic modeling of the cardiovascular system. The invention can obtain good three-dimensional reconstruction results of angiography, effectively assists the detection and operation navigation of cardiovascular diseases, and meets the clinical requirements.

An image scanner is provided with a super-resolution processing function for generating a high-resolution image using multiple pieces of low-resolution image data. The scanner device then measures an initial MTF value without performing the super-resolution processing, refers to an MTF properties table in which is registered a number of pieces of low-resolution image data for a scanner device of that type to achieve a pre-set target MTF, and registers the number of images that corresponds to the initial MTF value in the scanner, if that number is present, and registers an interpolated value if that number is not present. The image scanner then performs the super-resolution processing using the registered number of pieces of low-resolution image data, and outputs image data of a target resolution.

The invention discloses an electrocardio vector reconstruction method based on unsupervised learning. The method carries out the reconstruction of an electrocardio vector of an inputted standard 12-lead electrocardiogram through employing a neural network. In the training process, a method of firstly mapping the standard 12 lead to the electrocardiogram vector and then restoring the 12 lead electrocardiogram by using a projection method is used, so that the problem that the traditional method depends on the corresponding data of the 12 lead and the electrocardiogram vector is solved, the utilization efficiency of the data is obviously improved, and the data cost is reduced. During reconstruction, a neural network is used for recalculating a projection vector to perform reconstruction froman electrocardiogram vector to a 12-lead electrocardiogram, and a regularization term is used for constraining the projection vector in a final loss calculation module, so that the interpretability and the accuracy of a reconstruction process are ensured while the individual difference of the electrocardiogram is solved. The multi-order differential loss is used in a final loss calculation module,so that the problems of low frequency, such as baseline interference and the like, are avoided on the basis of ensuring morphological characteristics.