54 results about "Atlas data" patented technology

According to one embodiment there is provided a method of selecting a plurality of M atlases from among a larger group of N candidate atlases to form a multi-atlas data set to be used for computer automated segmentation of novel image data sets to mark objects of interest therein. A set of candidate atlases is used containing a reference image data set and segmentation data. Each of the candidate atlases is segmented against the others in a leave-one-out strategy, in which the candidate atlases are used as training data for each other. For each candidate atlas in turn, the following is carried out: registering; segmenting; computing an overlap; computing a value of the similarity measure for each of the registrations; and obtaining a set of regression parameters by performing a regression with the similarity measure being the independent variable and the overlap being the dependent variable. The M atlases are then selected from among all the N candidate atlases to form the multi-atlas data set, the M atlases being those atlases determined to collectively provide the highest aggregate overlap over all the training data image sets.

According to one embodiment there is provided a method of selecting a plurality of M atlases from among a larger group of N candidate atlases to form a multi-atlas data set to be used for computer automated segmentation of novel image data sets to mark objects of interest therein. A set of candidate atlases is used containing a reference image data set and segmentation data. Each of the candidate atlases is segmented against the others in a leave-one-out strategy, in which the candidate atlases are used as training data for each other. For each candidate atlas in turn, the following is carried out: registering; segmenting; computing an overlap; computing a value of the similarity measure for each of the registrations; and obtaining a set of regression parameters by performing a regression with the similarity measure being the independent variable and the overlap being the dependent variable. The M atlases are then selected from among all the N candidate atlases to form the multi-atlas data set, the M atlases being those atlases determined to collectively provide the highest aggregate overlap over all the training data image sets.

According to one embodiment there is provided a method of selecting a plurality of M atlases from among a larger group of N candidate atlases to form a multi-atlas data set to be used for computer automated segmentation of novel image data sets to mark objects of interest therein. A set of candidate atlases is used containing a reference image data set and segmentation data. Each of the candidate atlases is segmented against the others in a leave-one-out strategy, in which the candidate atlases are used as training data for each other. For each candidate atlas in turn, the following is carried out: registering; segmenting; computing an overlap; computing a value of the similarity measure for each of the registrations; and obtaining a set of regression parameters by performing a regression with the similarity measure being the independent variable and the overlap being the dependent variable.

An object of the present invention is to provide a medical image data alignment apparatus, method and program capable of aligning different kinds of image data on a tissue to be objected at high precision and in a short time. According to the present invention, a patient SPECT and a patient CT are aligned via atlas data. The atlas data is a standard of SPECT image data, and is created based on SPECT image data of a plurality of patients. The patient SPECT and the atlas data are aligned based on the correlation of image signal values thereof, and a first transformation matrix T1 is determined. The atlas data and the patient CT are aligned based on the coordinate information added to these data in advance, and a second transformation matrix T2 is determined. The patient SPECT and the patient CT are aligned using the first transformation matrix T1 and the second transformation matrix T2.

The invention discloses a cable partial discharge pattern recognition method and system which can be used for reducing the amount of data required for calculation, shortening recognition time, and effectively acquiring electric discharge characteristics of different partial electric discharges. Without losing characteristic parameters the method and system can be used for obtaining significant characteristic values from the subtle parts and improving classification recognition rates. The method comprises the following steps: data in m cycles is taken from each data group and superimposed in one cycle to form atlas data of each data group; 360 degrees in the cycle is subjected to phase window dividing operation in units of the same angle, and n equal-interval phase windows are obtained; thecharacteristic value of the atlas data in the cycle corresponding to each phase window is calculated, and a first characteristic value matrix is obtained; the obtained first characteristic value matrix is subjected to dimension reduction operation so as to obtain a second characteristic value matrix, characteristic values in the second characteristic value matrix are classified and identified viaa pattern identification classifier, and a classification result corresponding to each sample signal is obtained.

Disclosed herein are techniques for enhancing the accuracy of atlas-based auto-segmentation (ABAS) using an automated structure classifier that was trained using a machine learning algorithm. Also disclosed is a technique for training the automated structure classifier using atlas data applied to the machine learning algorithm.

The invention relates to a spindle turning error source tracing method based on shaft center orbit manifold learning. The method includes the following step that (1) two electrical vortex sensors are arranged on the periphery of a spindle at intervals and used for collecting spindle vibration signals; (2) the detected spindle vibration signals are processed to judge an operation state of the spindle; (3) the spindle vibration signals intersect at one point on the same plane, and a shaft center orbit is obtained after continuous sampling; (4) error separation is conducted on a spindle center orbit to obtain spindle actual rotation precision A; (5) a mapping function atlas data base Q:{f(i)=Qij|A} is obtained according to the spindle actual rotation precision A and a manifold sensitive characteristic Qij; and (6) if the spindle actual rotation precision A>=etaE, eta=0.8-1, the mapping function atlas data base Q is called, source tracing of spindle rotation errors is conducted, and corresponding faults are maintained; and if the spindle actual rotation precision A>=etaE, eta=0.6-0.8, source tracing analysis monitoring is conducted on the spindle rotation errors, wherein E is spindle rotation precision of a machine tool leaving a factory.

The invention discloses an automatic construction technology for predicting a next sentence model based on a BERT model. The method comprises test atlas data acquisition and natural language inferencemodel construction training and prediction, a test atlas data acquisition part can be connected with an atlas database to automatically acquire data with specified relations in all APPs related to acertain field. The invention relates to the technical field of natural language processing. According to the automatic construction technology for predicting a next sentence model based on a BERT model, through applying a natural language reasoning technology in deep learning to the field of APP testing, a node pair with a next sentence relationship in the graph database is automatically obtained,and the node pair is automatically processed and converted into training data required for predicting a next sentence of model. The BERT-based prediction next sentence model is used for realizing automatic reasoning and assisting in completing automatic construction of the map, so that the working efficiency is improved, and compared with other natural language reasoning models, the BERT-based prediction next sentence model has higher prediction accuracy.

The present invention discloses a method, a system and an apparatus for detecting a large-scale complex network community structure. The method comprises: abstracting a to-be-detected large-scale complex network as atlas data; using a multi-thread parallel sliding window model to carry out optimized storage on the abstracted atlas data; using a multi-thread parallel adaptive tag propagation algorithm to carry out tagged processing on the stored atlas data; and carrying out post-processing according to a tagged processing result and outputting a community structure detection result. The systemcomprises an atlas abstraction module, an optimized storage module, a tagged processing module and a post-processing module. The apparatus comprises a memory and a processor. According to the technical scheme of the present invention, time complexity is reduced and the execution efficiency is improved; the technical scheme of the present invention can also compute the large-scale atlas through anordinary personal computer, so that the cost is reduced; the technical scheme of the present invention can adaptively identify overlapping and non-overlapping communities, so that the community detection accuracy is improved; and the technical scheme of the present invention can be widely applied in the field of complex network service computing.

An object of the present invention is to provide a medical image data alignment apparatus, method and program capable of aligning different kinds of image data on a tissue to be objected at high precision and in a short time. According to the present invention, a patient SPECT and a patient CT are aligned via atlas data. The atlas data is a standard of SPECT image data, and is created based on SPECT image data of a plurality of patients. The patient SPECT and the atlas data are aligned based on the correlation of image signal values thereof, and a first transformation matrix T1 is determined. The atlas data and the patient CT are aligned based on the coordinate information added to these data in advance, and a second transformation matrix T2 is determined. The patient SPECT and the patient CT are aligned using the first transformation matrix T1 and the second transformation matrix T2.

The invention discloses a power equipment fault probability prediction method considering insulation defect categories and fault relevance. The method comprises the following steps: (1) collecting PRPS map data of power equipment and preprocessing the PRPS map data; (2) extracting partial discharge characteristics based on the pre-processed PRPS atlas data; (3) inputting the partial discharge characteristics into a trained convolutional neural network, and outputting a probability value P (Dk) that the power equipment has a certain type of insulation defects through the trained convolutional neural network; inputting the partial discharge characteristics into a trained long-short-term memory neural network, and enabling the trained long-short-term memory neural network to output the faultprobability P (F | Dk) of the power equipment under the condition of Dk; and (4) obtaining the final fault probability P (F) of the power equipment based on the following formula. In addition, the invention also discloses a power equipment fault probability prediction system.

1. A medical data processing method of determining information describing the probable position of a neural fibre in a patient's brain, the method comprising the following steps which are constituted to be executed by a computer: a) acquiring patient-specific medical image data describing the brain of the patient; b) acquiring atlas data defining an image-based model of a human brain; c) determining, based on the patient-specific medical image data and the atlas data, seed region data describing seed regions (A, B) in the patient-specific medical image data in which the ends of neural fibres of the patient's brain may be located; d) determining, based on the patient-specific medical image data and the seed region data, neural fibre tract data describing a plurality of potential tracts (T1, T2, T3) which a specific neural fibre may take through the patient's brain; e) determining, based on the atlas data and the neural fibre tract data, a figure of merit for each one of the potential tracts (T1, T2, T3).

A system and method are provided for enabling atlas registration in medical imaging, said atlas registration comprising matching a medical atlas 300, 302 to a medical image 320. The system and method may execute a Reinforcement Learning (RL) algorithm to learn a model for matching the medical atlas to the medical image, wherein said learning is on the basis of a reward function quantifying a degree of match between the medical atlas and the medical image. The state space of the RL algorithm may be determined on the basis of a set of features extracted from i) the atlas data and ii) the image data. As such, a model is obtained for medical atlas registration without the use, or with a reduced use, of heuristics. By using a machine learning based approach, the solution can easily be applied to different atlas matching problems, e.g., to different types of medical atlases and / or medical images.

The invention discloses a partial discharge on-line monitoring alarm confidence coefficient analysis method. The method includes the steps that partial discharge of electrical equipment is monitored on line to obtain real-time monitoring data; the real-time monitoring data are compared with fault atlas data to generate partial discharge alarm information; according to statistics of the type and the frequency of the partial discharge alarm information, corresponding confidence coefficients are set for the partial discharge alarm information; the confidence coefficients corresponding to the partial discharge alarm information are compared with a threshold value, if the confidence coefficients of the partial discharge alarm information are larger than or equal to the threshold value, the partial discharge alarm information is output, and if the confidence coefficients of the partial discharge alarm information are smaller than the threshold value, the partial discharge alarm information is abandoned. The invention correspondingly provides a partial discharge on-line monitoring alarm confidence coefficient analysis device and system. According to the embodiment, the accuracy and the reliability of partial discharge on-line monitoring alarming can be effectively improved, the invalid working amount is reduced, and the device maintaining cost is saved.

The invention discloses a system and a method for determining concentration of carbon dioxide in industrial smoke. The method comprises the following steps: preparing a series of simulated industrial smoke with known carbon dioxide standard gas concentration by virtue of a standard gas storage tank, a gas mass flow meter and a controller of the detection system; respectively determining absorption peak areas of carbon dioxide standard gases of different volume concentrations by virtue of an IGS gas infrared analysis instrument in the system, and establishing a standard fitting working curve according to data; and collecting and measuring to-be-determined industrial smoke, substituting a measurement result into the standard fitting working curve, so as to obtain the content of carbon dioxide gas in the industrial gas. According to the method provided by the invention, the original atlas data of the area of a quantitative absorption peak of carbon dioxide can be recorded, related data can be preserved for a long time, the method can meet the value traceability requirement of carbon emission data, the result is accurate, and the precision is high.

The invention discloses a method for classifying sensory substances based on olfactory brain waves and GS-SVM, which comprises the following steps of: S1, utilizing brain-computer interface system, i.e., the brain electric instrument, to acquire the electroencephalogram spectrum information of the subject; S2, preprocessing the acquired EEG spectrum data; S3, performing feature extraction on thepreprocessed atlas data based on the linear characteristic and the nonlinear characteristic analysis, 76-dimensional data including peak, mean, standard deviation, center value, center frequency, power sum and LZC complexity of alpha, beta, theta frequency bands are used as brain electrical characteristics in the study of brain electrical signals; S4, adopting a network format search support vector machine (GS-SVM) for pattern recognition. According to the method for classifying sensory substances based on olfactory brain waves and GS-SVM, the physiological morphology of the human brain information processing process in the product evaluation process is truly restored, which has extremely important significance in the fields of clinical medicine and cognitive science and can be widely usedin the sensory evaluation of substances, making the sensory evaluation process more concise, more standardized, precise and scientific.

An image data processing apparatus comprises a data receiving unit for receiving image data to be segmented, and an atlas selection unit for accessing a plurality of atlas data sets and selecting a subset of the atlas data sets for use in segmenting the image data, wherein the atlas selection unit is configured to select the subset of atlas data sets in dependence on the positions of one or more anatomical landmarks comprised in the plurality of atlas data sets.

The invention discloses a power equipment failure rate prediction method based on a convolutional neural network, and the method comprises a training step and a prediction step, and the training stepcomprises the steps: (1) collecting a case PRPS map of power equipment; (2) preprocessing the collected case PRPS atlas data; (3) constructing a first convolutional neural network module, and trainingthe first convolutional neural network module to enable the first convolutional neural network module to output a defect type corresponding to the case PRPS spectrum data; (4) constructing a data setof each defect type based on the defect type; (5) respectively constructing respective fault dichotomy sub-modules corresponding to the defect types, wherein each fault dichotomy sub-module is constructed based on a second convolutional neural network module; and training a second convolutional neural network to enable each fault binary classification sub-module to obtain a fault occurrence probability value based on the case PRPS map data, and outputting a judgment whether the power equipment has a fault or not.

The invention discloses a data producing, organizing, storing and accessing method of an electronic atlas system. The method comprises the steps of: acquiring grid pictures; performing naming; entering metadata; establishing a description folder, a legend folder, a thumbnail folder, a picture folder and an index file, and organizing production of an atlas; cutting a single map into map tiles; dividing map products into single maps, map pairs, series maps and atlas; uploading the map products to the cloud server to establish the atlas database index; downloading the map tile data into the innercard or directly pre-loading the data into the inner card; putting different map tile data into different wild cards; carrying out map inquiry and reading as required. The invention greatly improvesthe retrieval speed of map data search, reduces the storage pressure of mobile equipment, enables people to read map with or without a network, and realizes the purpose of dynamically opening and sharing the atlas data based on the background data center.

The invention discloses a model training method and system and electronic equipment, and the method comprises the steps: obtaining standard data and original data, building a first incidence relation between the standard data and the original data, obtaining a training sample, building a second incidence relation between the standard data and the reality data according to the semantic similarity between the standard data and the reality data, according to the first incidence relation and the second incidence relation, obtaining initial atlas data, putting the training sample into the initial atlas data, putting the initial atlas data into an atlas neural network model for N times of training, and obtaining N loss values, and taking the training model corresponding to the minimum loss value in the N loss values as the prediction model. Learning training is carried out on the initial map data through the method to obtain the prediction model, and the data element corresponding to the maximum loss value is screened out through the prediction model when the input fields are matched, so that the accuracy of a field association result is improved.

The invention provides a method for optimizing atlas resources in LayaIDE and a readable storage medium, and the method comprises the steps: integrating to-be-packaged atlas resources according to a space occupation optimization strategy, and generating atlas data and atlas resources; acquiring the atlas data, and converting the atlas data into a standard format supported by a Laya project; and integrating the converted atlas data and the atlas resources. The configurable size of the atlas in the LayaIDE can be realized, and the memory occupancy rate of the atlas can be remarkably reduced; furthermore, the method has the advantages of simplicity and convenience in operation, easiness in implementation and the like.

The invention discloses an atlas data processing method and device and electronic equipment, and the method comprises the steps: receiving initial atlas data sent by a user side, obtaining first data and second data corresponding to the initial atlas data, screening the first data based on a first preset condition, obtaining first effective data in the first data, and sending the first effective data to the user side; and screening the second data based on a second preset condition to obtain second valid data in the second data, and generating valid atlas data corresponding to the initial atlas data according to the first valid data and the second valid data. According to the method, the initial atlas data are screened through the first preset condition and the second preset condition, it can be ensured that the screened data are effective atlas data, and effective screening of the data stored in the atlas structural form is further achieved.

The invention provides a method for detecting metabolites of patients with diabetic nephropathy in treatment of a Gandi capsule, characterized in that the specific steps are as follows: S1. selectinga test group and a control group according to a ratio of 1:1; S2. periodically collecting a sample to be tested; S3 treating the sample in the S2, and performing UPLC-Q-TOF / MS separation analysis; S4,performing metabolic contour analysis on the UPLC-Q-TOF / MS atlas data obtained in S3 to obtain a data set; S5, constructing a partial least squares discriminant analysis model according to the data set obtained in S4; and S6, constructing an S-PLOT load diagram according to the model constructed in S5, calculating a VIP value, meanwhile obtaining a corresponding P value, and screening out a metabolic marker capable of distinguishing a blank group and a Gandhi capsule administration group by using it as a condition that the VIP value is greater than 1.0 and the P value is less than 0.05. By adoption of the method, the pharmacological mechanism of the Gandi capsule in the treatment of the diabetic nephropathy can be predicted.

Disclosed is a computer-implemented method of training a likelihood-based computational model for determining the position of an image representation of an annotated anatomical structure in a two-dimensional x-ray image, wherein the method encompasses inputting medical DRRs together with annotation to a machine learning algorithm to train the algorithm, i.e. to generate adapted leamable parameters of the machine learning model. The annotations may be derived from metadata associated with the DRRs or may be included in atlas data which is matched with the DRRs to establish a relation between the annotations included in the atlas data and the DRRs. The thus generated machine learning algorithm may then be used to analyse clinical or synthesized DRRs so as to appropriately add annotations to those DRRs and / or identify the position of an anatomical structure in those DRRs.

The invention discloses an atlas data reduction method based on PDF (Portable Document Format) file analysis. The method comprises the following steps: obtaining an atlas position range by analyzing a file; identifying and classifying data with different functions and relative coordinates according to position attributes of various related objects in the atlas; obtaining relative coordinates and absolute coordinates of a specific point in the atlas through a mutual relation between the data, and further obtaining a horizontal coordinate correction coefficient and a vertical coordinate correction coefficient corresponding to the relative coordinates and the absolute coordinates; and converting the obtained relative coordinate data to obtain absolute coordinate data for constructing the atlas, thereby realizing the reduction of the PDF atlas data. Herein, the map content in the PDF format is converted into data which reflects map characteristics, has a numerical value close to that of original data and can be operated and retrieved, so that the use of the map data is not limited by an original special system, a workstation and a working program, the convenience of exchange, query and comparison of the map data is improved, and the unified management of the data is facilitated.

The invention provides a method and device for obtaining atlas data, and the method comprises the steps: achieving a rule engine; utilizing the rule engine to realize a data extraction rule; and running the data extraction rule to extract data from a business system, and converting the extracted data into atlas data. The invention provides a method and a device for acquiring map data, which can reduce the workload of acquiring the map data.

The method includes the steps of obtaining atlas data in an atlas coordinate set from a computer-readable atlas of hepatic anatomical information including three orders of division and obtaining patient data in a patient coordinate set. The method further includes morphing atlas data from the atlas coordinate set to the patient coordinate set by performing a rigid registration between the at least one landmark identified in the patient coordinate set and the corresponding obtained atlas data in the atlas coordinate set and by performing a non-rigid registration of the three orders of division of the hepatic structure while maintaining the rigid registration of the at least one landmark.