103results about How to "Automatic recognition" patented technology

The invention belongs to the field of face recognition and discloses an advertisement delivery device and method based on face recognition, wherein the advertisement delivery device based on face recognition comprises at least one camera module, an image recognition device, a display device and an advertisement delivery switching device. The camera module is used for acquiring image information. The image recognition device comprises an image information receiving unit, a face recognition unit, a face feature recognition unit and a core operation unit, wherein the image recognition device receives image information acquired by the camera module, the core operation unit controls the face recognition unit and the face feature recognition unit to compute and analyze the image information, and the advertisement delivery switching device delivers a proper advertisement to the display device according to the computing and analyzing result. Through the advertisement delivery device and method based on face recognition in the invention, the advertisement value can be increased, and the effect of optimizing the advertisement benefit can be achieved.

The invention relates to the technical field of ATM intelligent monitoring systems. A module comprises an input module, a video intelligent monitoring server, a monitoring alarm management module and an output module. A video intelligent analysis system to which the video intelligent monitoring server is related particularly refers to the behaviors of automatically detecting an object, tracking the object, and recognizing and analyzing the object. The video intelligent monitoring server relates to an anomalous event analysis system which is combined with a front camera and an upside camera to monitor anomalous and dubious behaviors of objects round an ATM, and simultaneously obtain face information of a suspect and give an alarm in time when an anomalous situation happens.

The invention provides atrial fibrillation signal recognition method, device and equipment, wherein the method comprises: acquiring an electrocardiogram signal to be recognized; inputting the electrocardiogram signal to an atrial fibrillation signal recognition model pre-established, and putting results of atrial fibrillation signal recognition, wherein the atrial fibrillation signal recognition model is established by acquiring a set amount of electrocardiogram sample signals and corresponding identifier information that includes normal signal and atrial fibrillation signal identifier information, using SMOTE (synthetic minority over-sampling technique) to balance atrial fibrillation signals according to the amount of normal signals, establishing a network structure of multiple convolutional neural networks each of which is provided with a specific receptive field for recognizing the atrial fibrillation signal of corresponding granular size, inputting the normal signals and the balanced atrial fibrillation signals to the network structure, and training. The atrial fibrillation signal recognition method, device and equipment can provide improved reliability and accuracy of atrial fibrillation signal recognition.

The embodiment of the invention provides a method and device for recognizing the class of social contact short texts and a method and device for training classification models. The method for recognizing the class of the social contact short texts includes the steps that social contact short text data are acquired; text characteristic data are extracted from the social contact short text data; with the text characteristic data as input, first class information of the social contact short text data is acquired from at least two trained short text classification models; second class information of the social contact short text data is determined according to the acquired first class information of the social contact short text data. By means of the method and device for recognizing the class of the social contact short texts and the method and device for training the classification models, the class information of the social contact short texts can be automatically and accurately recognized, so that the effect and accuracy of classifying lots of social contact short texts are improved, and the method and device for recognizing the class of the social contact short texts and the method and device for training the classification models are widely applied to various short text analysis scenes and improve network experience of users.

An automatic identification method for ocean phytoplankton, comprising following steps: (1) making grape feature of known ocean phytoplankton be corresponding to its description and forming a data base of phytoplankton to store in computer memory in advance; (2) collecting grape feature of identifying phytoplankton with digital microscope and sending into computer; (3) comparing collected grape feature with that of in data base, and outputting description of known phytoplankton when the osculation reach a regulated value. The invention combines digital microscope with computer system to identify ocean phytoplankton automatically, fast and accurately.

The invention relates to an editing distance, word frequency and word vector based entity relation recognition method which comprises the following steps: 1) acquiring text data; 2) using word segmentation software to perform word segmentation and counting the number of each word in the text; 3) adjusting the dimensions of word vectors, window size and conducting word vector training; 4) using well-trained word vectors to converge words; and 5) calculating the editing distance of the entity obtained in the fourth step, and combining the word frequency obtained in the first step to obtain the alias or the abbreviation of a given entity; based on the editing distance and the step 4, calculating the d [i, j] value of the given entity word to other entities and combining step 1 to obtain the number of each entity word; and obtaining G (X) through weighted averaging, namely, obtaining first n possible abbreviations of a given entity word and finally deciding the alias or abbreviation of the given entity word through strength of the relationship by their weight.

The embodiment of the invention discloses a data processing method and device, and the method comprises the steps: obtaining all sessions between a target protection object and a plurality of communication objects when detecting that the target protection object is in an attacked state; obtaining a message set associated with all sessions, and determining session characteristic parameters of eachsession according to the message set; inputting the session characteristic parameters of each session into an identification model, and determining a session type of each session based on the identification model; determining a communication object type of each communication object in the plurality of communication objects according to the session type of each session; and determining the communication object of the illegal type as an illegal communication object, adding the illegal communication object to the blacklist, and discarding the session corresponding to the illegal communication object. By adopting the embodiment of the invention, the security of the communication information can be improved, so that the stability of a communication system is ensured, and meanwhile, the type range of the protected client can be expanded.

The invention discloses a working condition identification method for an indicator diagram of a beam pumping unit based on integrated learning, and belongs to the technical field of oil and gas production. The machine learning is performed by firstly substituting the geometric features, the moment features and the Fourier descriptors of the first preset partial indicator diagram in each of the indicator clusters under each working condition into at least two machine learning models; at least two sets of classifier models are obtained by training; the geometric features, the moment features andthe Fourier descriptors of the second preset partial indicator diagram are substituted into each set of classifier models; at least two classification results of each indicator diagram are trained; and at least two classification results of each indicator diagram and corresponding labels are substituted into the multinomial logistic regression model for integrated learning, and the final classifier model is obtained. When it is necessary to judge the working condition of the indicator diagram of the to-be-tested beam pumping unit, the geometric features, the moment features and the Fourier descriptors are substituted into the final classifier model, which can automatically and accurately identify the working condition of the indicator diagram.

The invention discloses a quick coal and rock recognition method based on DCT low-frequency component characteristics. According to the method, coal and rock image characteristic information is described through low-frequency components after DCT of images, and a coal and rock classifier is established through a learning vector quantitative method; in the recognition process, image characteristic vectors of the images to be recognized are extracted through a method similar to that of training images, the characteristic vectors of the images are input in the classifier, and recognition is conducted through the ratio of the characteristic vectors of the images to be recognized in the classifier. According to the method, the images in different degrees of illuminance and different viewpoints of coal and rock are used as training samples, the influence of illuminance and imaging viewpoint change is small, the recognition rate is high, and stability is good.

The embodiment of the invention relates to the technical field of medical image processing, in particular to a liver boundary identification method and system. The method comprises the following steps: obtaining liver tissue information to be identified; adopting an image processing technology or a signal processing technology to identify a liver tissue boundary in the liver tissue information according to the characteristics of a liver tissue corresponding to the liver tissue information and the characteristics of the liver tissue boundary; and outputting the position information of the identified liver tissue boundary. When the method is adopted, the liver tissue boundary can be automatically identified, and liver boundary identification efficiency can be improved so as to realize the automatic positioning of the liver boundary.

The invention relates to the technical field of artificial intelligence, and provides a sensitive information identification method, which comprises the following steps of obtaining a to-be-identifiedfield from a database table of a database; if at least one piece of information in the to-be-identified field comprises a Chinese character string and if the Chinese character string comprises a character string meeting a first preset rule, determining whether the character string is matched with at least one character string template in a preset corpus or not; if the character string is not matched with any character string template in a preset corpus, determining a hash value corresponding to the character string; determining a target character string of which the hash value is matched withthe hash value corresponding to the character string in a preset corpus; determining the similarity between the target character string and the character string; and if the similarity between the target character string and the character string exceeds a preset value, determining that the to-be-identified field comprises a sensitive field. The invention further provides a device, electronic equipment and a storage medium. The sensitive field can be automatically identified.

The invention discloses a sanitation robot capable of automatically dumping garbage and a control method of the robot. The robot comprises a chassis, a dustbin body, a lifting frame, a lifting rope and a moving mechanism driven by a drive motor; the dustbin body is arranged on the chassis, two bottom plates capable of being opened and closed and a servo motor for controlling the bottom plates to be opened and closed are arranged at the bottom of the dustbin body, the middle of the lifting frame is transversely bent, the lower end of the lifting frame is connected with the chassis through a rotating disc, the rotating disc is controlled by the servo motor to rotate, the free end and the bending position of the upper portion of the lifting frame are both provided with pulleys, one end of the lifting rope is connected with the dustbin body, and the other end of the lifting rope winds around the two pulleys and is connected with an output shaft of a winch mounted at the lower end of the lifting frame. The moving mechanism is arranged below the chassis. The robot can replace sanitation worker to automatically finish garbage cleaning and dumping, garbage on the corners can be cleaned, meanwhile, the robot can be suitable for cleaning different kinds of garbage, and the cleaning efficiency is improved.

The invention discloses a method for recognizing coal and rock on the basis of co-occurrence features of image blocks. The method includes densely extracting the image blocks, extracting key image blocks by the aid of clustering algorithms, labeling coal and rock images by the aid of vectors of the key image blocks, computing co-occurrence matrixes of the labeled images, and extracting energy, contrast and inverse difference moment and entropy of the co-occurrence matrixes to form features of the images; comparing the to-be-recognized images to various modes and determining that the most similar modes are categories of the to-be-recognized images. The features of each sample image represent the corresponding mode of the coal or the rock. The method has the advantages of little influence of illumination and imaging view point change, high recognition rate and good stability.

The embodiment of the invention discloses a preset type number recognition method and device. The method comprises the steps that starting from a node corresponding to a first type number in an operating relation chain, the node serves as a starting node, adjacent other nodes are searched according to the point direction of branches of the node till leaf nodes, a father node of the leaf nodes is traced, adjacent other nodes of the father node are searched, and whether the attribute of the searched nodes meets the specific condition or not is judged in the process; the steps are repeatedly executed till all nodes in the operating relation chain are accessed, and a first judgment result is obtained; the nodes meeting the specific condition in the first judgment result are screened according to the preset screening condition, and the preset type number is obtained. According to the preset type number recognition method and device, the preset type number can be automatically recognized.

The invention relates to a determination method and a measurement system for three spontaneous combustion zones of a goaf of a coal mine. By virtue of a network concentration controller arranged in anunderground roadway and wireless gas concentration sensors arranged in an underground goaf, one wireless gas concentration sensor sends location information and gas concentration measurement information of all the wireless gas concentration sensors to the network concentration controller by virtue of an ad hoc network, the network concentration controller determines the three spontaneous combustion zones of the goaf according to the location information and gas concentration measurement information of each wireless gas concentration sensor, the network concentration controller is utilized forrealizing dynamic drawing and display of three spontaneous combustion zones of a goaf of a coal working face, a visual data basis is provided for performing reasonable operations, such as working face drawing speed design, roadway layout design, gas extraction design, fire alarm and emergency rescue, on a mine, informatization level of the mine can be greatly improved, and fine management on themine is realized.

The invention discloses a coal rock recognition method based on relativity measurement learning. According to the method, a new relativity measurement function is learnt from a training sample set in a monitoring mode to measure the relativity of coal and rock image samples, so that the relativity measurement value of samples in the same type is larger and larger, the relativity measurement value between samples in different types is smaller and smaller, and the classification rate of unknown samples is increased. The method includes an image preprocessing process, a training process and a recognition process. A preprocessing module is used for simple preprocessing of collected coal and rock images to obtain the training sample set. A training module is used for learning the optimal coal rock classification effect relativity measurement function from the training sample set. A recognition module is used for performing measurement classification by means of the optimal relativity measurement function. By means of the method, images of coal and rock under different illuminances and different viewpoints serve as the training samples, the method is little influenced by illuminance and imaging viewpoint changes, the recognition rate is high, and stability is good.

The embodiment of the invention provides a risk identification method and device, electronic equipment and a storage medium. The risk identification method comprises the steps that parameter data of at least one parameter type corresponding to a target complaint work order is obtained; and according to the parameter data of the at least one parameter type, a risk identification model is utilized to identify a target risk level of the target complaint work order, and the risk identification model is obtained through training according to the parameter data of the at least one parameter type corresponding to the sample complaint work order and the risk level. According to the technical scheme provided by the embodiment of the invention, the identification accuracy of the security risk level of the complaint work order can be improved.

The invention belongs to the technical field of computer network information security control, and discloses a computer network information security control system and method. The computer network information security control system comprises a network information acquisition module, a network information transmission module, a main control module, a network information identification module, an encryption module, an intrusion detection module, an alarm module, a cloud storage module and a display module. After the information to be forwarded is received through the network information transmission module, the information to be forwarded is forwarded to a next-level information receiver, so that in the multi-stage transmission process of the information, each stage of information receivercan acquire the source of the information, the transmission path of the information and the relationship between the information sender and the information receiver in each stage, at the same time, whether the to-be-identified network information is credible or not can be determined through the network information identification module background, that is, whether the to-be-identified network information is credible or not is determined by utilizing the similarity, and accordingly the rumor can be automatically and effectively identified.

The invention relates to the field of artificial intelligence, and discloses a lung feature recognition method, a device, computer equipment and a storage medium. The method comprises the steps of obtaining to-be-recognized data which comprises a to-be-recognized lung image and to-be-recognized lung text description; performing lung image feature extraction through the lung image recognition modelto generate a lung image feature vector and an image recognition result, and performing lung text feature extraction through the lung text recognition model to generate a lung text feature vector anda text recognition result; fusing the lung image feature vector and the lung text feature vector by using an attention mechanism through a lung fusion recognition model, and extracting image text fusion features for recognition to obtain a fusion recognition result; and obtaining a lung feature recognition result through voting. According to the invention, the lung features can be accurately recognized, and the recognition accuracy and reliability are improved. The method is suitable for the field of intelligent medical treatment, and can further promote the construction of intelligent cities.

The invention relates to a method and a terminal for identifying the industry to which an enterprise belongs, and belongs to the field of data processing. The method comprises the steps of training afirst deep learning model according to a plurality of bill samples; wherein the first deep learning model is used for obtaining an industry category to which an enterprise belongs according to the bill information; wherein each bill sample comprises a bill and an industry category corresponding to the bill; acquiring bills of an enterprise to obtain a bill set; wherein the bill set comprises billscorresponding to the main business operation range of the enterprise; calling a preset first deep learning model to identify the bill set to obtain a first industry category table; wherein the industry category table comprises more than one record; each record includes a name of an industry category and a probability that the enterprise belongs to the industry category. The accuracy of identifying the industry to which the enterprise belongs by the intelligent terminal is improved.