899results about How to "Effective classification" patented technology

An improved method for efficiently retrieving data associated with linked network locations utilizing a computer network is disclosed. First, data is retrieved from a network location which has a plurality of links to related network locations. Next, in response to a single user input, the plurality of links and their network addresses are automatically classified into a folder having common subject matter. A user can later efficiently retrieve data associated with the classified network locations utilizing the common category of network addresses.

Scanning of computer files for malware uses a classifying technique to classify an input file as a clean file or a dirty file. The parameters of the classifying technique are derived to train the classification on a corpus of reference files including clean files known to be free of malware and dirty files known to contain malware. The classification is performed using a representation of the files in a feature space defined by a set of predetermined features for respective file formats, the features being a predetermined value or range of values for one or more data fields of given meanings. The representation of a file is derived by determining the file format, parsing the file on the basis of the structure of data fields in the determined file format to identify the data fields and their meaning, and determining, on the basis of the identified data fields, which of the set of predetermined features are present.

A method for classifying electronically posted documents includes receiving two posted documents and generating corresponding metadata summaries for each, wherein each of the metadata summaries includes at least one sub-tree structure. The structures of the two summary sub-trees within the respective metadata summaries are subsequently compared. If the two summary sub-trees are different, the two documents are deemed distinct. If the two summary sub-trees are the same, attribute values and text content of the metadata summaries are compared over a portion of the metadata summaries. If the compared attribute values and text content are determined to be the same, the documents are deemed duplicative.

The invention belongs to the field of data processing technology and discloses a text sentiment classification method facing Chinese Web comments. The text sentiment classification method includes a training process and a classification process. The training process includes the steps of carrying out training text preprocessing, carrying out feature selecting, carrying out vectorization representation of a text and obtaining a training classifier. The classification process includes the steps of carrying out test text preprocessing, carrying out feature selecting, utilizing the classifier to classify and outputting a classification result. On the basis of an original document classification method, document frequency (DF) and information gain (IG) are used and a sentiment dictionary of negative words, degree adverbs and dynamic sentiment words are built to distinguish sentiment tendency of Chinese feature words, select feature words, calculate a feature weight value and build a feature vector. Moreover, a NaiveBayes classification algorithm is used for training to obtain the classifier, carrying out sentiment classification on the text, providing effective data mining for users and then carrying out analysis processing.

A system and method of providing an autocomplete recommended word, which classify a recommended word list according to indexes of various languages, store the recommended word list for each index, extract a corresponding autocomplete recommended word according to a user query and a setting mode which is received from a user's web browser, provide the user with the corresponding autocomplete recommended word, and thereby may propose a suitable recommended word according to the user query.

Disclosed is a circular index structure. The contents feature analyzer analyzes the contents and stores information about virtual locations to be located on the indexes of the respective contents, and information on the physical indexes for displaying the corresponding contents, in a memory. The contents selector extracts the information on the physical indexes using the information stored in the memory according to the user's request, and displays the contents. Therefore, the user can easily select desired various contents having various channels and categories such as the digital televisions and the web more intellectually, and accordingly, the user's confusion caused by the overflow of information is prevented and the user can quickly access the contents.

The invention enables the digital management of large scale image databases, to efficiently classify and index image data independent of language. Complex processing requirements are required only on reduced and operably small subsets of the entire collection, thereby effectively scaling large document collections. Embodiments of the present invention provide image-based classification and retrieval of documents based on image recognition, e.g., signatures, logos, stamps, or word-spotting; in documents within real time for large datasets such as in the millions of documents.

A system and method provide for efficient classification of long strings of data, such as network messages. The system, which may be a classification engine for use in a network device, is configured to include one or more stages having one or more banks of ternary content addressable memories (TCAMs). The TCAM banks of each stage are organized into one or more groups and each group processes the network messages for a different set of the network device's ports. The system further includes at least one memory resource that is shared by the TCAM banks of all groups. That is, the parallel banks of TCAMs operate in a distributed fashion while the shared memory resource operates in a centralized fashion. Accordingly, the system can process network messages at high speed while minimizing the number of required components.

A method of classifying a test subject comprises collecting imaging data for a plurality of training subjects, control subjects and a test subject. An intensity volume of interest (VOI) and a spatial VOI are selected from said imaging data. Training intensity data and spatial data are calculated for the intensity and spatial VOI. A statistical model can then be created based on the training intensity data and training spatial data to provide a universe of subjects. Control intensity data and spatial data are also calculated for the intensity and spatial VOI. A classifier can then be built dividing the universe into at least two regions. The test subject data can then be applied to the classifier to provide a determination of whether said test subject falls within the first region or the second region. The condition can be a neurological disease state such as temporal lobe epilepsy or Alzheimer's dementia.

The invention discloses a fabric defect detection method based on a depth neural network. The method comprises the following steps: (1), an image acquisition system is built to acquire an image; (2), the image is segmented into experimental samples, fabric sample image data are enhanced at the same time, and a fabric image after enhancement serves as a training sample; (3), a depth neural network is designed; (4), parameters are set, the depth neural network is initialized, the training sample is fed to the depth neural network for training, and after network training is completed, the network model is saved; and (5), an inputted new fabric sample is fed to the network model for detection. According to the fabric defect detection method based on the depth neural network provided by the invention, with a convolutional neural network as a core, feature extraction is performed by a convolutional layer, a pooling layer retains effective features and reduces the amount of calculation, and a full connection layer is used for classification. A mini-batch gradient descent method is used for optimization, the generalization ability is enhanced through L2 regularization, defect recognition is carried out through determining the corresponding position of the maximum component outputted by a classifier, effects are shown in Figure 4, Actual presents the actual category of the sample, and Pred presents the predicted category of the sample.

The invention discloses a method for identifying a disturbance event in a distributed type optical fiber pipeline security early-warning system. When the disturbance event exists, wavelet de-noising is conducted on two routes of sampling signals. The characteristic values of one sampling signal where wavelet de-noising is conducted are extracted, wherein the characteristic values include the vibration fragment length, the time domain energy, the k-order original point distance, the k-order center distance, the skewness, the kurtosis and low frequency wavelet coefficient energy Ej, obtained through wavelet decomposition, of all layers, and j ranges from 1 to 7. The thirteen extracted characteristic values are sent to a decision tree classification device, and the type of the disturbance event is obtained through the decision tree classification device. Man-machine interaction incremental learning is achieved by changing the type of the disturbance event stored in a database under the condition that a new type of the disturbance event appears or the type, obtained through the decision tree classification device, of the disturbance event is wrong, and online training is conducted on the decision tree classification device according to the modified type of the disturbance event. By means of the method, the type of the disturbance event can be accurately obtained.

The implementation of the invention discloses an adaptive football video summary generation method, receiving the football video streaming using shot boundary detection method based on the average sliding window frames to do shot boundary detection in football video streams to get the video sets; use video classification method based on sub lens region to classify the video sets; do wonderful video detection in the classified video sets and output the detected wonderful video as video summary. The invention also discloses an adaptive football video summary generation device, which can improve the efficiency and accuracy of football video summary generation.

Disclosed are a mobile terminal and a method for controlling a 3D multi-angle view thereof. Multi-angle views of a 3D image captured by a plurality of cameras are displayed, switched (converted) and selected according to a touch input. And, a selected angle view of the 3D image undergoes a capturing or a recording. While appreciating different-angle views of a 3D image (broadcast image) captured by a plurality of cameras, a desired angle view of the 3D image may be played, recorded and stored with using a user interface. This may enhance a user's convenience.

The invention relates to a radar-simulation-image-based human body motion classification method of a convolution neural network. The method comprises: step one, establishing a time-frequency image data set containing a variety of human motions; step two, enhancing radar time-frequency image data; step three, establishing a convolution neural network model; to be specific, using a handwriting recognition network LeNet as a basis, introducing a modified linear unit ReLU to replace an original Sigmoid activation function as an activation function of the convolution network based on three convolution layers, two pool layers and two fully connection layers, adding a pool layer and reducing a fully connection layer to form a convolution neural network structure including three convolution layers, three pool layers and one fully connection layers, and adjusting an interlamination structure and an in-layer structure of the network and training parameters to realize a good classification effect; and training the convolution neural network model.

The present invention relates to an online advertisement classified pushing method and system based on a consumer behavior data analysis and classification technology. Compared with the prior art, the online advertisement classified pushing method and system overcome the defect that potential customers cannot be mined to carry out network online advertisement pushing. The online advertisement classified pushing method comprises the following steps of: carrying out data collection and preprocessing, i.e. collecting behavior data of consumers from online mobile terminals, establishing a data pool, carrying out preprocessing operation on the data in the data pool and providing data support for subsequent data analysis and modeling; aiming at the behavior data of the consumers, carrying out modeling, i.e. establishing a topic model facing the behavior data of the consumers so as to mine relations between the consumers and online advertisement categories as well as a purchase time period; and aiming at the consumers, carrying out effective classification and aiming at different consumer categories, pushing the corresponding types of advertisements online. According to the present invention, by collecting the behavior data of the consumers on various mobile terminals, carrying out analysis and modeling on behaviors of the consumers and mining consumption habits of different consumers, effective classification of the consumers is implemented.

The invention discloses an automatic classification method of an electrocardiosignal ST band. The automatic classification method is characterized by comprising the following steps: S1. acquiring an electrocardiosignal wave form of a human body and pretreating the electrocardiosignal wave form; S2. performing characteristic point detection to the pretreated electrocardiosignal wave form; S3. based on the characteristic point detection in the step S2, determining the wave form of the electrocardiosignal ST band, and acquiring the characteristic parameters on the wave form of the electrocardiosignal ST band so as to establish a to-be-classified characteristic input matrix; S4. classifying the wave form of the electrocardiosignal ST band into a training sample and a testing sample, and establishing a classifier model based on the training sample; and S5. inputting the testing sample into the classifier model for testing, and completing the final classification by combining decision fusion. The invention further discloses an automatic classification system and device of the electrocardiosignal ST band. By establishing the classifier model and decision fusion by using a nerve network method, calculation can be effectively reduced, the time cost can be decreased, the classification precision of the ST band can be improved, and the classification is easier.

A method for efficiently categorizing images on a computer system is disclosed. A series of related images that are to be categorized are ordered by time of capture, and category levels for input of category information by a user are displayed. The category levels include a highest-category level and a lowest-category level, where the highest-category level has a low frequency of data change between the series of related images, and the lowest-category level has a high frequency of data change between the series of images. A first image is then categorized by allowing the user to enter highest-category level data and lowest-category level data. A next image in the series is then categorized by leaving the highest-category level data unchanged, and automatically selecting the lowest-category level data for reentry by the user, thereby eliminating the need for the user to reenter the highest-category level data.

Methods and apparatus are provided for classifying data packets in a data processing device according to a set of processing rules, wherein, for each of a predetermined group of data items in each packet, each rule defines a rule range indicating a range of possible values of the corresponding data item for which that rule applies. The method comprises for each data packet: (a) performing a preliminary test for at least one data item in said group, the preliminary test comprising testing the value of the data item in the packet for a match with any of a predetermined set of frequently-occurring values for that data item, each frequently-occurring value being associated with a predetermined indicator, and, if a match is obtained, selecting for the data item the indicator associated with the matching frequently-occurring value; (b) performing an item search for any data item in the group for which no match is obtained in a preliminary test, the item search for a said data item comprising selecting a range identifier corresponding to the value of the data item from a predetermined set of range identifiers for that data item, the set of range identifiers indicating, for all possible values of the data item, which of the rule ranges corresponding to the data item in the rule set a value intersects, wherein the indicator associated with a said frequently occurring value for a data item is the range identifier corresponding to that value of the data item; and (c) identifying, based on the selected range identifiers for all data items in the group, at least one rule of any rules applicable to the data packet.

The invention discloses a waste and old mixed plastic recovery and separation device and method, and belongs to the field of resource recycling of waste and old mixed plastic. The waste and old mixed plastic recovery and separation device comprises a crushing device and a melting, sorting and recovering device, wherein a material feeding mechanism, a crushing mechanism and a material discharging mechanism are arranged in a crushing box body of the crushing device; a heating device, a screening mechanism and a plastic molding mechanism are arranged in a melting box body of the melting, sorting and recovering device; the heating mechanism is used for heating crushed granular materials to a molten state in a staged way; the screening mechanism is used for separating the materials in the molten state and un-melted granular materials; the plastic molding mechanism is placed on the bottom of the melting box body and recycling the separated materials in the molten state. The recovery and separation device can effectively classify the waste and old plastic generated in daily life and industrial production; different plastics are separated and are made into a single plastic which is relatively pure, so that the resource utilization of the waste and old plastic is realized; the post-processing process is omitted; energy resources are saved; the operation is simple; the automation degree is high; the sorting degree is high; the sorting range is wide.

The present invention relates to a process for classifying the panel MURA in the module inspection of a liquid crystal display panel. The invention includes a MURA area logical operation process (S5) of photographing the liquid crystal display panel from different angles of visibility, performing an image processing for detecting a MURA area for a group of images taken and then performing an image logical operation process, an upper-level classification process (S6) of classifying the shape of MURA, and a lower-level classification process of classifying the panel MURA by combining an upper-level classification with other parameters. With this invention, the panel MURA is correctly detected without setting up the complicate parameters. This invention contributes to labor saving at the final inspection step for manufacturing and leads to quality assurance and higher reliability of the liquid crystal display panel in the liquid crystal display panel manufacturing field.

A defect is efficiently and effectively classified by accurately determining the state of overlap between a design layout pattern and the defect. This leads to simple identification of a systematic defect. A defective image obtained through defect inspection or review of a semiconductor device is automatically pattern-matched with design layout data. A defect is superimposed on a design layout pattern for at least one layer of a target layer, a layer immediately above the target layer, and a layer immediately below the target layer. The state of overlap of the defect is determined as within the pattern, over the pattern, or outside the pattern, and the defect is automatically classified.

The invention provides a data parallel processing method. The data parallel processing method comprises the following steps that 1, a main management node receives data and acquires the incidence relation of the data; 2, the main management node calculates allocatable GPUs and GPU work loads of work computing nodes; 3, the main management node partitions the data and distributes the partitioned data to all the work computing nodes; 4, the work computing nodes perform parallel processing on the received data and transmit processing results back to the main management node; 5, the main management node merges the results and then outputs the results. The data parallel processing method has the following advantages that a master-slave architectural pattern is adopted to be used for high-performance large-scale data parallel processing, operation stage partition is performed on specific operations converted by application programs according to DNA feature modeling, node granularity grade operation deployment is performed according to a partition result, and the execution efficiency of a parallel task of data flow in a single node is improved by adopting a thread parallel optimization mechanism and fully utilizing multiple computing kernels.

The classification system of a network device includes a cache in which a mapping between predefined characteristics of TCP / IP packets and associated actions are stored in response to the first “Frequent Flyer” packet in of a session. Selected characteristics from subsequent received packets of that session are correlated with the predefined characteristics and the stored actions are applied to the received packets if the selected characteristics and the predefined characteristics match, thus reducing the processing required for subsequent packets. The packets selected for caching may be data packets. For mismatched characteristics, the full packet search of the classification system is used to determine the action to apply to the received packet.

A statistical system and method for generating patterns and performing online handwriting recognition based on those patterns. A plurality of predetermined patterns may be generated by performing feature extraction operations on one or more character samples utilizing a Gabor filter. An online handwritten character may be acquired. The online handwritten character may be pre-processed. One or more feature extraction operations, utilizing a Gabor filter, may be performed on the online handwritten character to produce a feature vector. One or more patterns may be generated, using a statistical algorithm, for the online handwritten character, based on the feature vector. The online handwritten character may be statistically classified based on a comparison between the one or more patterns generated for the online handwritten character and the plurality of predetermined patterns.

The invention discloses a fault classification method based on a one-dimensional multi-path convolutional neural network, and belongs to the technical field of industrial process monitoring. Accordingto the method, a traditional two-dimensional convolutional neural network is improved; multiple paths of parallel one-dimensional convolutional neural networks are applied in the variable direction;the cross correlation among variables is deconstructed, the variables are independently subjected to convolution and pooling to extract time sequence feature information, the extracted features are more diversified, the robustness is higher, the defect that a traditional two-dimensional convolutional neural network is sensitive to the prior arrangement order of the variables in input data is overcome, and the method is more suitable for complex and high-order industrial process data; experiments show that the fault classification model obtained based on the one-dimensional multi-path convolutional neural network training provided by the invention can effectively perform fault classification of industrial process data, and has higher generalization capability compared with a common model.

The scheme of the invention discloses a vehicle assistance driving system based on the infrared thermal imaging and millimeter wave radar technology. The main functions include pedestrian and vehicledetection, image and sound fusion and collision early warning in driving of the vehicle. The early warning system uses the following technical characteristics: the infrared camera and the millimeter wave radar specialized for vehicle-mounted assistance driving, the data acquired by the radar, such as the distance, the direction and the speed and the images acquired by the infrared camera are fusedfor screening and determining the pedestrians or the vehicles, the pedestrian or vehicle characteristics are extracted through the algorithm, finally the result is displayed on the visual interface and multidimensional early warning information including images, numbers and sound is provided.

The invention discloses a method for selecting and rapidly comparing robust features of iris images. The method comprises: S1, adopting regional multipole sequencing measurement features to express iris texture; S2, selecting and fusing iris features by taking similarity as guidance; and S3, rapidly comparing iris characteristic templates on the basis of a hierarchical classifier. The method uses a feature-selecting method taking similarity as the guidance to select regional measurement features and construct a combined classifier, and adopts a hierarchical classifier structure to optimize and combine the obtained combined classifier again so as to rapidly compare the iris characteristic templates. The method automatically learns the obtained iris regional sequencing measurement features, and has the advantages of greatly raising the speed of comparing the iris characteristic templates and effectively solving the problem of applying iris recognition to large-scale databases, along with good robustness and high recognition precision. The method can be widely used in a large number of application systems using iris for identity recognition and security protection.

The invention provides a handwritten Chinese character layout evaluation method. The method comprises the following steps: (1), pretreatment; (2), performing evaluation to the position layout of each stroke in a handwritten Chinese character; (3), performing layout evaluation to all the parts in the handwritten Chinese character by part layout evaluation; (4), performing evaluation to the integral form of the handwritten Chinese character by integral font evaluation; and (5), combining the results of evaluation in the steps (2), (3) and (4) to obtain the evaluation conclusion. The invention adopts the five steps of pretreatment, stroke layout evaluation, part layout evaluation, integral form evaluation and overall evaluation, and obtains the evaluation conclusion by measuring nine kinds of indexes, the conclusion takes the character reasoning as a primary way and the grading as an auxiliary way, and the writing quality of the users is effectively classified by grade according to the index value.

The invention discloses a three-dimensional point cloud model classification method based on convolution neural network, includes selecting Princeton ModelNet to generate training set and data set from training data and test data by selecting required number of models from official website according to ModelNet 10 and ModelNet 40 respectively, selecting training data and test data from official website according to Princeton ModelNet, selecting Princeton ModelNet to generate training set and data set according to model Net 10 and ModelNet 40 respectively, and selecting Princeton ModelNet to generate training data and test data. 2, carry out feature analysis on that point cloud model and constructing a classification framework; S3, ordering the point cloud; S4, two-dimensional visualizing the ordered point cloud data; S5, Constructing CNN network for two-dimensional point cloud image. The invention applies the CNN in the image field directly to the classification of the three-dimensional point cloud model for the first time, 93.97% and 89.75% classification accuracy were obtained on ModelNet 10 and ModelNet 40 respectively, Experimental results show that it is feasible to classify 3D point cloud model by using CNN in image domain. PCI2CNN proposed in this paper can capture 3D feature information of point cloud model effectively and is suitable for classification of 3D point cloud model.