109 results about "Spatial learning" patented technology

The invention relates to a cross-media retrieval method based on deep learning and consistent expression spatial learning. By starting with two methods including feature selection and the similarity estimation of two highly-isomerous feature spaces, the invention puts forward the cross-media retrieval method capable of improving multimedia retrieval accuracy to a large extent by aiming at the cross-media information of two modalities including an images and a text. The method disclosed by the invention is a multimedia information mutual retrieval method which aims at two modalities including the image and the text, and cross-media retrieval accuracy is improved to a large extent. In a model which is put forward by the invention, a regulated vector inner product is adopted as a similarity metric algorithm, the directions of the feature vectors of two different modalities are considered, the influence of an index dimension is eliminated after centralization is carried out, an average value of elements is subtracted from each element in the vectors, the correlation of the two vectors subjected to average value removal is calculated, and accurate similarity can be obtained through calculation.

A method and apparatus for automatic detection and labeling of 3D spinal geometry is disclosed. Cervical, thoracic, and lumbar spine regions are detected in a 3D image. Intervertebral disk candidates are detected in each of the spine regions using iterative marginal space learning (MSL). Using a global probabilistic spine model, a separate one of the intervertebral disk candidates is selected for each of a plurality of labeled intervertebral disk locations.

The invention discloses a cross-modal retrieval algorithm based on mixed hypergraph learning in subspace, based on the cross-modal public subspace learning of canonical correlation analysis. The similarities inside and among modals are calculated by a mapping of the public subspace. A mixed relation matrix is calculated by the similarities inside and among the modals. A mixed hypergraph model is built by the extraction of the relation matrix. And the cross-modal retrieval and sample sequencing are conducted by adoption of hypergraph learning finally. With aiming at cross-modal heterogeneous variations and high order relation among samples, the algorithm instance applies the hypergraph model combined with the cross-modal public subspace learning to the cross-modal retrieval, so that the model is capable of considering the similarity among the modals and the similarity inside the modals simultaneously and giving consideration to the high order relation among a plurality of the samples meanwhile, improving the final precision ratio and the final recall ratio of the cross-modal retrieval. The algorithm is capable of effectively improving the performance of the cross-modal retrieval and greatly enhancing the precision ratio and the recall ratio of the cross-modal retrieval.

The invention provides a method and a system for re-identifying a pedestrian based on view angle self-adaptive subspace algorithm. With the method, a rectangular image only comprising a single pedestrian or a target rectangular frame cut from an original video image is used as an input image; a feature vector is extracted from the input image; a data set is divided into a training data set and a testing data set; a transformation matrix is acquired on the training data set by learning the view angle self-adaptive subspace learning algorithm; and the transformation matrix acquired on the testing data set by learning the view angle self-adaptive subspace learning algorithm is utilized to calculate the distance and re-identify the pedestrian. According to the invention, the fact that different cameras have different properties is taken into account; and different transformations are used to compensate the transformation properties of different cameras, thus the optimum transformation relationship between each pair of the cameras can be acquired more flexibly through learning; and the features under different cameras after the cameras are transformed can be closer to the ideal feature distribution.

A method and apparatus for detecting multiple anatomical landmarks in a 3D volume. A first anatomical landmark is detected in a 3D volume using marginal space learning (MSL). Locations of remaining anatomical landmarks are estimated in the 3D volume based on the detected first anatomical landmark using a learned geometric model relating the anatomical landmarks. Each of the remaining anatomical landmarks is then detected using MSL in a portion of the 3D volume constrained based on the estimated location of each remaining landmark. This method can be used to detect the anatomical landmarks of the crista galli (CG), tip of the occipital bone (OB), anterior of the corpus callosum (ACC), and posterior of the corpus callosum (PCC) in a brain magnetic resonance imaging (MRI) volume.

The invention discloses a multi-visual angle gait recognition method and a system based on higher-order tensor subspace learning, which belong to the field of intelligent recognition. A gait video is acquired from multiple representational angles, and a gait sequence image is obtained through framing interception; background extraction, background subtraction and binary processing are carried out on the gait sequence image respectively, black and white visual effects are presented, and a contour sequence under the multiple visual angles is obtained; the contour sequence is converted to tensor data; a higher-order discriminant tensor subspace analysis algorithm based on graph embedding obtained after expanding DTSA on the basis of multilinear discriminant analysis and a graph embedding principle is used for carrying out dimension reduction and feature extraction on the tensor data; and according to the extracted and obtained multi-visual angle gait features, the gait features are subjected to similarity measurement, and a recognition result is obtained. The method is simple, the cost is low, person identity authority detection and disguised person identity authentication can be automatically carried out on a particular place, and safety protection on the monitored place and identity authentication in multiple conditions can be effectively improved.

The invention provides a figure regular hyperspectral image band selection method based on subspace learning. The method comprises the concrete realization steps that (1) a hyperspectral image data matrix is inputted; (2) the data matrix is normalized; (3) a figure regular band similarity matrix is calculated; (4) a figure regular band similarity diagonal matrix is calculated; (5) a reconstruction matrix is constructed; (6) the reconstruction matrix is initialized; (7) the number of times of iteration is set; (8) a subspace band selection matrix is calculated; (9) whether the current number of times of iteration is greater than the maximum number of times of iteration is judged, and the step (10) is performed if the judgment result is yes, or one time of current iteration is added and the step (8) is performed; (10) the subspace band selection matrix is outputted; and (11) a subspace data matrix is constructed. A learning mechanism is provided and spectral space geometric structure information is utilized so that the accuracy of band selection can be enhanced.

The invention discloses a gait recognition method and system in combination with subspace learning and tensor neural network and belongs to the field of intelligent recognition. The method includes the following steps: acquiring gait data, obtaining a gait data set, and processing the gait data set to obtain a silhouette set, and further obtaining a gait energy diagram; taking 80% of the silhouettes as a training set and conducting dimensionality reduction on the training set, taking the rest 20% of the silhouettes as testing set data and testing the result of the training, then extracting features from the gait energy diagram and a data tensor neural network module which is subject to dimensionality reduction, then classifying the features through a support vector machine which acts as a classifier, finally comparing the training set and the result of the test set to obtain the result from identifying and verifying the identity of passenger. According to the invention, the method is easy to implement and has low cost. The method can automatically conduct identity authorization and fake identity verification in a certain venue, and effectively increase the effectiveness of identity verification for security and protection and other conditions in a monitored venue.

The invention provides a stupid-rat-maze-type intelligence rehabilitation training device, comprising multiple layers of training cages, the multiple layers of training cages are arranged up and down in a laminating way, adjacent training cages are mutually nested and connected and are communicated by virtue of a ladder; a maze is arranged in each training cage, the mazes comprise relaxation mazes, plug board mazes, elevated anxiety mazes and escape hole mazes, and the mazes can be combined and arranged at will. In the invention, natural behavioral expression of conditioned response of a stupid rate to construct a voluntary activity maze learning memory training device, improvement of stupid rat behaviour function expression is promoted by virtue of a learning memory maze training space which is fresh, changeable, interesting, comfortable and ready to activities, spatial learning capability and memory capability are improved, search capability and retelling capability of an animal canbe more accurately and objectively reflected, the stupid-rat-maze-type intelligent rehabilitation training device provided by the invention is beneficial to improvement and evaluation on learning andmemory functions of senile dementia, and rehabilitation process is sped up, thus the stupid-rat-maze-type intelligence rehabilitation training device provided by the invention has great significance.

The invention relates to a multi-space fusion learning environment construction method and device. The method comprises the following steps: pre-defining a plurality of spaces included in a multi-space fusion learning environment and each space learning environment construction parameter; constructing each spatial semantic network model; collecting a learning subject associated event and a learning environment associated event, and constructing a learning subject-oriented multi-space semantic hierarchy consistency data fusion model; constructing an integrated body model based on a learning scene; and setting the learning environment construction parameters for the learning main body according to the integrated tool body model. According to the invention, a multi-space fusion learning environment can be dynamically constructed for a learning subject in a personalized manner.

The present invention relates to a method and device for multi-scale characteristic extraction based on ECG. The method and device firstly obtain an ECG signal identification unit based on a lead ECGsignal interception and performs multi-scale decomposition of the ECG signal identification unit to construct an ECG multi-scale space; then ECG multi-scale space signals in the ECG multi-scale spaceare subjected to multi-scale characteristic extraction by a preset convolutional neural network; the ECG signal identification unit comprises at least one cardiac cycle band which is more conducive for the convolutional neural network to conduct ECG variation characteristic and spatial characteristic learning of myocardial infarction, and extracts variation characteristics from a deeper level fromECG signals through the preset convolutional neural network, and the variation characteristics have a strong ability to discriminate diseases, obtain spatial characteristics related to locations of the diseases based on spatial learning ability of the convolutional neural network, and provide important practical reference value for doctors to predict the locations of the myocardial infarction.

The invention discloses a local subspace learning-based multi-view clustering method. The method comprises the steps of performing characteristic extraction on data in different views and generating a corresponding kernel matrix; obtaining a unified kernel matrix by using a local subspace learning-based multi-view method; performing mapping by using a Laplace mapping algorithm; and performing clustering on expressions in a low-dimensional space by adopting a Kmeans algorithm to obtain a result. The multi-view clustering method is realized and the clustering effect is improved.

The invention relates to a domain self-adaption technology in the field of machine learning, and provides a Grassmann manifold domain self-adaption method based on symmetric matrix space subspace learning. In order to reduce the difference between the probability distribution of the source domain data and the target domain data, the method comprises the following steps of firstly establishing mapping from a Grassmann manifold to a symmetric matrix space, then mapping the Grassmann manifold matrix data of the source domain and the target domain to the symmetric matrix space, and constructing asubspace in the symmetric matrix space; subjecting the projection of the original data in the subspace to a mean value similarity criterion; establishing an objective function of the subspace learning, optimizing and solving the objective function to obtain the objective subspace, matching the projection probability distribution of the original data in the objective subspace on the objective subspace, that is, achieving the domain self-adaption of the original data through twice transformation from the Grubman manifold to the symmetric matrix space and then from the symmetric matrix space to the subspace of the symmetric matrix space.

The invention discloses a semi-supervised image clustering subspace learning algorithm based on local linear regression. Firstly, a local linear regression model is used for predicting a coordinate of a training sample in a clustering subspace, a local prediction error between a predicted value and a true value is obtained, and then a minimized objective function of a total predicted error is obtained; then according to two constrain conditions of inter-class dispersion maximization and inner-class dispersion minimization, and a marked sample and an unmarked sample are used for calculating an inter-class dispersion matrix and a total dispersion matrix; and finally, the inter-class dispersion matrix and the total dispersion matrix are blended in the minimized objective function of the total predicted error to obtain an objective function for solving clustering subspace, and function solving is performed through generalized characteristic root to obtain the optimal clustering subspace. The semi-supervised image clustering subspace learning algorithm based on the local linear regression makes full use of the marked sample, the unmarked sample and a local adjacent relation in a training data set to obtain good clustering results.

A face recognition method based on multi-view collaborative complete discriminant subspace learning is provided. The method comprises the following steps: (1) using an objective function based on Cauchy loss and Fisher discriminant analysis to obtain complete feature representation as shown in the specification of the number as shown in the specification of training samples in a potential completesubspace, the number as shown in the specification of view generation functions as shown in the specification, and the number as shown in the specification of non-negative collaborative learning weights as shown in the specification; (2) given the non-convex nature of the objective function, obtaining two solutions as shown in the specification of the objective function by using the alternate solution method; (3) based on the solved view generation functions as shown in the specification and the non-negative collaborative learning weights as shown in the specification, solving complete feature representation of test samples in the complete discriminant subspace; and (4) based on the Euclidean distance between the test sample and the training sample in the complete discriminant subspace, classifying the test samples by using a nearest neighbor classifier. Compared with the existing multi-view face recognition method, the method provided by the present invention can more effectively fuse multi-view information and mine discriminant information, and is an effective multi-view face recognition method.

A human body activity recognition method based on grouping residual joint spatial learning comprises the following steps: step 1, collecting human, object and environment signals by using various sensors, grouping, aligning and slicing single-channel data based on a sliding window, and constructing a two-dimensional activity data subset; step 2, building a grouping residual convolutional neural network, and constructing a joint space loss function optimization network model by utilizing a center loss function and a cross entropy loss function in order to extract a feature map of a two-dimensional activity data subset; and step 3, training a multi-classification support vector machine by utilizing the extracted two-dimensional features to realize a human body activity classification task based on the feature map. According to the invention, fine human body activities can be identified; the inter-class distance of the extracted spatial features is increased in combination with a joint spatial loss function, and the intra-class distance is reduced; based on the spatial feature map of the human body activity data, a multi-classification support vector machine is combined to carry outclassification learning on the feature map, and the accuracy of human body activity classification is improved.

Provided are methods for reducing the amount or activity of C90RF72 RNA, and in certain instances of reducing the amount of C90RF72 protein, in an animal. Such methods are useful to prevent or ameliorate at least one symptom of a neurodegenerative disease. Such symptoms include anxiety, reduced spatial learning, and memory loss. Such neurodegenerative diseases include amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FTD), corticalbasal degeneration syndrome (CBD), atypical Parkinsonian syndrome, and olivopontocerellar degeneration (OPCD).

The invention provides a cross-media retrieval method based on subspace learning and semi-supervised regularization, which is characterized by comprising the following steps: Step 1, establishing a multimedia database, collecting multimedia original data, extracting features of multimedia data, and storing feature vectors of the multimedia data and the original data; Step 2, acquiring a projectionmatrix of different media types, defining an optimal target function, solving the optimal target function by utilizing an iterative method, and projecting the feature vectors of the multimedia data to a public space; Step 3, carrying out cross-media retrieval, extracting features of media data submitted by a user, projecting feature vectors of the media data into the public space, calculating a similarity between the projected vectors and other vectors in the public space, and returning the media data corresponding to first k feature vectors with the largest similarity with other vectors in the public space. The cross-media retrieval method provided by the invention generates a more accurate retrieval result.

The invention discloses a medical composite for preventing and curing Alzheimer's disease (AD). The composite comprises the following raw materials in parts by weight: 10-40 parts of hyperin, 20-80 parts of icariin and 5-20 parts of Schisandrin B. Experiments of rat place navigation and space exploration prove that the medical composite of the invention can increase the ability of spatial learning and memory of AD rat model, namely reduce the escape latency and scouting distance and obviously increase the percent of scouting time in quadrant of original platform to total time and the percent of scouting distance in quadrant of original platform to total distance in space exploration experiments. Therefore, the extract composite can be used for preventing and curing AD.

The invention discloses novel application of forsythin belonging to active components of natural medicinal plants, and particularly relates to application of forsythin as the only active ingredient to preparation of a medicine for improving a cognitive function and treating Alzheimer's diseases. Animal and cell research results indicate that the forsythin has the exact function of improving the cognitive function and can remarkably improve the spatial learning and memory ability of animals in an aging and Alzheimer's disease model, the level of a center monoamine neurotransmitter of an old-aged animal is increased, the genetic expression of presenilin 2 in a rat brain of an Alzheimer's disease model is reduced, and the survival rate of neuroblastoma cells induced by beta-amyloid protein is increased. The forsythin is matched with relevant auxiliary materials, a health care product or a medicine for purposefully improving the cognitive function and treating the Alzheimer's diseases can be prepared by a conventional preparation method, and the health care product and the medicine can be used for delaying and improving relevant diseases of cognitive hypofunction such as the course of the Alzheimer's diseases, enhancing the healthy quality and the life quality of old people and realizing a healthy aging social value.

The embodiments of the invention disclose a coupled spatial learning-based scene character recognition method. The method includes the following steps that: inputted scene character images are preprocessed, so that trained scene character images are obtained; recognition feature extraction is performed on the trained scene character images, so that a spatial dictionary can be obtained; the spatial dictionary is utilized to perform spatial coding on the recognition features of corresponding images, so that corresponding spatial coding vectors can be obtained; maximization extraction is performed on the spatial coding vectors, so that feature vectors can be obtained; a linear support vector machine is utilized to perform training based on the feature vectors, so that a scene character recognition classification model is obtained; and the feature vectors of a test scene character image is obtained and is inputted into the scene character recognition classification model, so that a scene character recognition result can be obtained. According to the coupled spatial learning-based scene character recognition method of the invention, the spatial dictionary is crated and is utilized to perform spatial coding, and therefore, the textual information of a space can be effectively integrated into the feature vectors, so that spatial information can be effectively mined, and therefore, the correct rate of scene character recognition can be improved.

The invention provides an application of polygala alkaline hydrolysis product composition in preparation of an anti-senile dementia medicine. The polygala alkaline hydrolysis product composition comprises the following main components by weight percentage: 25-38% of tenuifolin, 10 to 25% of fallaxsaponin A, 12 to 25% of 3, 4, 5-trimethoxycinnamic acid (TMCA) and 5-15% of p-methoxycinnamic acid (PMCA). The polygala alkaline hydrolysis product composition can obviously improve learning and memory impairment of a C57 mouse caused by hippocampal injection of Alpha Beta 25-35 and can obviously improve the spatial learning and memory capability of a C57 mouse model suffering from the learning and memory impairment caused by scopolamine, with an effect superior to that of a positive medicine, as well as can obviously improve the spatial learning and memory capability and the passive avoiding capability of a rapidly-aging C57 mouse model suffering from the learning and memory impairment.

The invention discloses an unsupervised feature selection method which combines mutual information and fractal dimension to solve the problem that the information fusion performance of subspace learning algorithm is degraded due to redundancy and uncorrelated features contained in multi-dimensional original features. Firstly, an original feature extraction method is used to extract the multi-dimensional feature information of the product, and the multi-dimensional feature information of the product in time domain, frequency domain and time-frequency domain is obtained. Secondly, based on the definition of mutual information and considering the redundancy and correlation among multi-dimensional features, the feature importance is sorted to obtain the sorted multi-dimensional feature set. Then, based on fractal theory, the feature subsets of sorted multidimensional feature sets are selected by fractal dimension feature subset evaluation criteria, and the optimal feature subsets are obtained. Finally, the subspace learning method is used to reduce the dimension of the optimal feature subset, and the product comprehensive features are obtained. On the basis of comprehensively considering feature redundancy and correlation, the method removes features with small correlation degree with extra multi-dimensional original feature set and large redundancy degree, improves information fusion performance of subspace learning method, and simultaneously obtains product comprehensive features.

The invention discloses a medical dataset characteristic dimension reduction method based on subspace learning. The method comprises the following steps of constructing an original high-dimension datamatrix X and a label column according to a to-be-analyzed medical dataset; constructing a most optimized target function, and solving a Lagrangean function thereof; according to the original high-dimension data matrix and the label column, calculating global discriminating information and local discriminating information; iterating and solving a conversion matrix Q until the target function is convergent or reaches a highest cycling number-of-times, thereby obtaining a dimension-reduced data matrix; training a model according to the calculated conversion matrix, calculating an AUC value evaluation dimension reduction matrix and classification accuracy. Compared with an existing characteristic dimension reduction method of the medical dataset, the method according to the invention is advantageous in that the local discriminating information and the global discriminating information of data are simultaneously used for performing dimension reduction; the method is suitable for the characteristic dimension reduction problem in a common scale, and relatively high classification accuracy is realized when the characteristic scale of the data is far higher than the sample scale.

The invention discloses the application of forsythiaside in preparing medicine promoting nerve cell proliferation and protecting nerve cells; the application of forsythiaside in preparing medicine improving passive avoidance response and impaired spatial learning and memory and enhancing learning and memory; the application of forsythiaside in preparing medicine increasing the total protein content in cerebral cortex tissue; the application of forsythiaside in preparing medicine reducing the activity of acetylcholin esterase in cerebral cortex and hippocampus brain tissue; the application of forsythiaside in preparing medicine improving the activity of superoxide dismutase in cerebral cortex and hippocampus brain tissue, and reducing the malonaldehyde content and the monoamine oxidase activity in cerebral cortex and hippocampus brain tissue; the application of forsythiaside in preparing medicine reducing the content of inflammatory factor TNF-alpha and IL-6 in the brain tissue of cerebral cortex; and the application of forsythiaside in preparing medicine treating mild cognitive impairment, especially senile dementia.

The invention relates to a new medicinal part of scutellaria baicalensis, and discloses a scutellaria baicalensis flower extract as well as a preparation method thereof and application thereof. The preparation method of the scutellaria baicalensis flower extract comprises the steps: taking a certain amount of scutellaria baicalensis flowers, using 8 to 10 times the mass of 50 to 90% ethanol as a solvent for reflux extraction 2 to 3 times correspondingly, each time lasts for 1 to 3 hours, combining the extract, concentrating under reduced pressure and freeze-drying to obtain powder. The pharmacodynamic experiments proves that the scutellaria baicalensis flower extract prepared by the invention can improve spatial learning and memory abilities and oxidative damage of aging rats, and can be applied in the preparation of anti-aging medicines and health care products.

The invention provides an unknown disease category identification method and device based on centralized space learning. The method comprises the steps of training an initial model through known category samples, initializing a known category space, and mapping the known category samples to a hyperspherical surface of a hidden space; training a generative adversarial network; generating an unknownanchor point; generating an unknown image corresponding to the unknown anchor point based on the unknown anchor point and the adversarial network; taking the unknown image as an unknown category sample, combining the unknown category sample with a known category sample, and adjusting a known category space to obtain a trained model; and obtaining a new sample, extracting features of the new sample based on the trained model, calculating the distance from the features of the new sample to each known category prototype point, if the distance is less than a preset threshold, determining that thenew sample belongs to a category corresponding to the known category prototype point, and if the distance is greater than or equal to the preset threshold, determining that the new sample belongs toan unknown category.

The invention discloses traditional Chinese medicine composition for treating the Alzheimer's disease as well as a preparation method and an application of the traditional Chinese medicine composition. The traditional Chinese medicine composition comprises milkvetch roots, Chinese angelica, red peony roots and prince's-feather fruits. Research indicates that the traditional Chinese medicine composition can improve the spatial learning ability and the spatial exploration ability of Kunming mice suffering from the Alzheimer's disease, increase visceral indexes of brains and spleens of the mice suffering from the Alzheimer's disease, reduce the permeability of Alzheimer's disease blood brain barriers, decrease Alzheimer's disease brain tissue hippocampal Abeta, IL-6 and IL-12 protein expression levels and have no obvious effect on weight of Alzheimer's disease models, namely, the Kunming mice. The results indicate that the traditional Chinese medicine composition is good in treatment effect on the Alzheimer's disease, so that the traditional Chinese medicine composition has the broad application prospect in the field of treatment of the Alzheimer's disease.

The invention discloses a vegetable oil composition with an effect of preventing and treating Alzheimer's disease as well as a preparation method thereof. The vegetable oil composition is prepared from the following raw materials in parts by weight: 1-20 parts of gardenia oil, 5-50 parts of jujube seed oil, 1-20 parts of ginkgo oil, 1-20 parts of Chinese magnoliavine fruit oil, 10-80 parts of flaxseed oil and 5-50 parts of safflower oil. The vegetable oil composition provided by the invention is prepared by screening and optimizing by virtue of a lot of experiences, and is also prepared by adopting vegetable oil which is prepared from natural traditional medicinal and edible plants with multiple medicinal healthcare effects by virtue of a modern preparation process and has a healthcare effect; and experimental results show that the vegetable oil composition provided by the invention can improve spatial learning and memory abilities of a senile dementia model mouse and shorten the escape latency and the scouting distance of the senile dementia model mouse, can reduce the oxidation damage degree of brain tissues of the mouse and inhibit the activity of acetylcholine esterase in the brain tissues at the same time, has very good prevention and treatment effects on the Alzheimer's disease, and can be used for preventing and treating the symptoms of amnesia, insomnia, dysphoria and the like of the Alzheimer's disease.

The invention provides a combination medicament for preventing and treating dementia and / or cognitive impairment. Particularly, the combination medicament comprises a) non-steroidal anti-inflammatory drugs or pharmaceutical salts thereof of effective dose for prevention and / treatment, b) statins compounds or pharmaceutical salts thereof of effective dose for prevention and / treatment, and c) optional pharmaceutically acceptable carrier or excipient. According to the combination medicament, the object cognition capability, passive avoidance response capability, active avoidance response capability and spatial learning and memorizing capability of model mice can be effectively increased and improved, therefore, the combination medicament can be used for preventing or treatment dementia and / or cognitive impairment and other diseases.