173results about How to "Convenience to follow" patented technology

The invention discloses a converged network lane line detection method based on an attention mechanism and terminal equipment. The method comprises the following steps: synchronously acquiring a natural image and point cloud data of a road surface; inputting the natural image and the point cloud data into a pre-established and trained fusion network, and outputting a lane detection result, whereinthe fusion network performs information fusion processing on the point cloud data and the natural image by adopting an adding time sequence frame and an attention mechanism. According to the invention, continuous frames are used to improve the performance of the detection network so as to deal with complex situations such as mark loss and vehicle occlusion; and by splicing the low-dimensional features and the high-dimensional features through Skip Connection, image detail information which is continuously lost along with the increase of the network depth is made up, and the image is reduced by using Decoder to obtain a final result. The fusion network greatly improves the detection performance of the lane lines and can perform detection in various complex environments.

The invention relates to a preschool education interaction cloud platform system which comprises at least one server, a plurality of parent end systems, a plurality of teacher end systems and a plurality of kindergarten leader end systems. The parent end systems, the teacher end systems and the kindergarten leader end systems communicate with the servers and are sequentially controlled by background management accounts, parent end users, teacher end users and kindergarten leader end users. An information exchange platform is provided among teachers, children and parents and particularly comprises kindergarten leader ends, teacher ends and parent ends, the preschool education interaction cloud platform system has motivation, communication and understanding functions of circles, comment, growth record, class albums and the like, the parents and the teachers interact by means of little red flowers, ranking lists, activities, reward points and the like, and the parents deeply merge into preschool education and seamlessly experience growth together with the children.

The invention discloses a Yellow River ice semantic segmentation method based on a multi-attention mechanism double-flow fusion network. The method is used for solving the technical problem that an existing Yellow River ice detection method is poor in accuracy. According to the technical scheme, firstly, data sets are collected and labeled, the labeled data sets are divided into a training data set and a test data set, then a segmentation network structure is constructed, the network comprises shallow branches and deep branches, and a channel attention module is added to the deep branch, a position attention module is added to the shallow branch, the fusion module is used for fusing the shallow branches and the deep branches, the data in the training set is added into the network in batches, the constructed neural network is trained by adopting cross entropy loss and an RMSprop optimizer, and finally, a to-be-tested image is input and a test is carried out by using the trained model. According to the method, multi-level and multi-scale feature fusion can be selectively carried out, context information is captured based on an attention mechanism, a feature map with higher resolutionis obtained, and a better segmentation effect is obtained.

The invention discloses an intelligent home system comprising a central controller, an electric appliance control system, a damp-proof system, an information receiving module, a video collecting system, a crawler detecting system, an input circuit and an alarm apparatus, wherein the central controller is used for receiving a user demand instruction and information uploaded by each branch control system and controlling the operation of corresponding branch control systems, and is communicated with a cloud server; the electric appliance control system is used for controlling the state of each electric appliance in home; the damp-proof system is used for detecting the humidity and the temperature of the home; the information receiving module is used for receiving operation short messages sent by a user through a mobile terminal; the video collecting system is used for collecting home states; the crawler detecting system is used for detecting whether crawlers exist in corners of rooms according to the change of a magnetic field; the central controller is connected with the input circuit; and the central controller is further connected with the alarm apparatus. According to the intelligent home system disclosed by the invention, the user can conveniently pay close attention to the home states at any time when going out so that the intelligent home system is of convenience for the user, has wide applicability, can also achieve an active alarm effect; and through the intelligent home system, the user has excellent life experience and higher living quality at home.

The embodiment of the invention discloses a forum searching method and a forum searching device. The method comprises the following steps of: acquiring keywords for searching articles in a forum, and generating search results which are not sequenced according to the keywords; determining the attention information of the articles in the search results which are not sequenced according to the search results which are not sequenced; and sequencing the search results which are not sequenced according to the attention information, and sending the search results which are sequenced to a client. By the invention, the attention of the articles can be reflected in the search results, and the articles which are valuable can be displayed in the top search results.

The invention discloses a method for displaying advertising messages, which comprises the following steps of: analyzing the service record of online service, and acquiring the service theme of the online service; acquiring the advertising message related to the service theme according to the service theme of the online service; and when a user uses the online service, displaying the advertising message related to the service theme to the user interface window of the online service. The method acquires the service theme of the online service by analyzing the service record of the online service, and when the user uses the online service, the advertising message related to the service theme is displayed to the user interface window of the online service, thus maintaining the consistence of the displayed advertisement and the current interest of the user, greatly avoiding the disgust of the user to the displayed advertisement, enhancing advertisement effect, and improving the click rate and exposure rate of the advertisement. The invention simultaneously discloses a device and a system applying the method.

The invention belongs to the technical field of CT imaging, and discloses a deep learning method for suppressing noise and artifacts in a low-dose CT image and facilitating later accurate medical diagnosis and analysis. The specific technical scheme is as follows: the double-attention generative adversarial network for noise reduction and artifact removal of the low-dose CT image is used for extracting non-uniform and irregular noise features and tissue texture features with complex distribution respectively; a Res2Net discriminator network for multi-scale feature extraction is designed, so that the discrimination capability of the discriminator is improved, and the stability and robustness of adversarial training are enhanced; a multi-description loss function combining artifact attentionloss, artifact consistency loss, structure constraint loss, adversarial loss and pixel-level L1 loss is designed to further improve the function of each sub-network. According to the method, the phenomenon of under-noise reduction or over-noise reduction caused by high similarity of noise artifacts and tissue structure distribution is avoided.

The disclosure relates to a display method, device and system of a direct broadcasting room. The method comprises: a user selection instruction is obtained, wherein the user selection instruction carries an identifier of a target audience member that is any one of the audience in a direct broadcasting room at a direct broadcasting terminal; and the identifier of the target audience member is sent to a server side, so that the server side sends target audience member special display information to all audience terminals in the direct broadcasting room. On the basis of the technical scheme, the direct broadcasting terminal obtains the user selection instruction and sends the identifier of the target audience member to the server side, so that the server side sends target audience member special display information to all audience terminals in the direct broadcasting room; and the audience terminals can display head image information of the target audience member specially at a direct broadcasting interface based on the target audience member special display information, so that the users can pay attention to the target audience member conveniently. Therefore, accuracy and convenience of audience sharing in direct broadcasting are improved and the using experience of the user is enhanced.

The invention discloses a rural informatization service platform based on information customization push technology and an operating method thereof, belonging to the technical field of internet platform. The rural informatization service platform is composed of an information push module, a supply and demand information module, and an expert interacting module; the information push module comprises a subscription module, a synchronization module, a recommendation module and a read module, and has the functions of completing individualized information customization, local server synchronous remote information, intelligent recommendation and information presentation; the supply and demand information module comprises a n information browsing module, an information releasing module, an information reminding module and an information subscribing module, and has the functions of completing supply and demand information browsing, searching, releasing, customizing and reminding; and the expert interacting module provides a platform for farmers and agricultural experts to communicate, comprises a farmer message leaving module, an expert reply module and a message reminding module, and has the functions of completing farmer-inquiring-expert, messaging expert for consulting, expert reply and message reminding. The invention is designed to avoid complicated operations in search for internet massive information and ensure information to be effective in real time; and a user can obtain the most interesting information with convenience.

A computer-based education system monitors brain activity of the student to produce an “engagement” signal that may be compared against a dynamic threshold to identify a plurality of points demarcating periods of declining attention. These points are used to trigger modifications to the presentation of the educational program to promote student retention of the present information.

The present invention provides a control method and device for a smart device. The control method and device are applied to the smart device, wherein the smart device establishes a connection with a cloud server in advance. The method includes the following steps that: a control instruction sent by the cloud server is received, wherein the control instruction is generated by the cloud server according to an operation instruction and a communication protocol between the smart device and the cloud server, wherein the operation instruction is sent by the mobile terminal and aims at the smart device; the control instruction is analyzed, so that an analysis result can be obtained; and corresponding operation is performed according to the analysis result. Thus, the problem of operating a refrigerator can be solved.

The invention relates to an image description text generation method based on a generative adversarial network. The method comprises the following steps: 1) constructing an encoder for realizing feature extraction of an image; 2) performing word embedding on the text, and constructing a decoder for generating an image description text; 3) pre-training a generator composed of an encoder and a decoder according to maximum likelihood estimation; 4) constructing a discriminator based on a convolutional neural network and training the discriminator; 5) jointly training a generator and a discriminator; and 6) inputting the test image data of the description text to be generated into the trained generator, and outputting the generated description text. Compared with the prior art, the method has the advantages that the objective evaluation score of the generated text is improved, the interpretability is good, and the diversity is high.

The invention discloses a restaurant kitchen personnel behavior identification method based on an attention mechanism, and aims to effectively identify human body behaviors in restaurant kitchen videos and realize kitchen monitoring management. The method comprises the following steps: acquiring working videos of kitchen workers by using a camera, constructing a database of which the videos comprise clear actions, performing frame cutting on the database, and extracting optical flow images; inputting the video frames into a spatial stream network based on a spatial attention mechanism to obtain spatial features; inputting the optical flow image into a time flow network based on a long-short-term memory network to obtain a time sequence feature; and respectively inputting the space and timesequence features into a classifier to obtain classification scores, and carrying out score fusion to complete kitchen personnel behavior identification. According to the method, a spatial attentionmechanism is added, so that the model pays more attention to more important points in space; and the long-short-term memory network better reserves the time sequence information in the video, so thatthe kitchen personnel behavior recognition accuracy is improved.

The embodiment of the invention provides a method for updating a city list of weather forecast and a handheld mobile terminal. The function of updating of the city list of weather forecast is realized by a call record mode or / and a message record mode in the handheld mobile terminal; the method comprises the detailed steps of: carrying out sorting to home cities of phone numbers of contacts by counting the amount of call record or / and message record corresponding to home cities of phone numbers of contacts of all the contacts in the handheld mobile terminal; obtaining the home city list of phone numbers of corresponding contacts; and according to the list, generating an updated city list of weather forecast. The implementation of the embodiment of the invention overcomes the operations of manually carrying out adding and sorting and the like of cities concerned by a user one by one in the prior art, realizes the function of automatically updating the city list of weather forecast and further leads the operations to be more simple, convenient, flexible and humanized.

The invention provides an intelligent safeguard baby protective cover. The intelligent safeguard baby protective cover comprises intelligent security protection equipment, a shield, a soft curtain, an external power supply and mobile equipment software, wherein the intelligent security protection equipment is the core component of the intelligent safeguard baby protective cover, at first, can purify air in the shield, and then monitors environmental data (including air quality, PM 2.5 concentration, ultraviolet intensity, temperature, humidity, oxygen content, carbon dioxide content and negative oxygen-ion concentration) by various built-in sensors; the shield and the soft curtain have the radiation-proof function and are connected by magic tapes, zippers and the like; the shield is connected with a baby carriage or a crib by magic tapes, zippers, clips and the like. The intelligent safeguard baby protective cover provides comprehensive protection for the environments of babies, and is complete in function, safe, reliable, fully intelligent to control, convenient to operate and low in maintenance cost.

The invention discloses a low-cost NB-IOT (Narrow Band Internet of Things) wireless intelligent lock, which comprises a controller, a safety lock, a battery and a communication module, wherein the controller is electrically connected with a biological recognition module, an NFC (Near Field Communication) module, a keyboard module and a storage module; the safety lock comprises a sound and light warning device, a lock body picking detection module and a lock body driving and lock body detection module; the battery is electrically connected with the battery volume detection module; the battery is electrically connected with the controller, the safety lock and the communication module; the communication module also comprises an NB-IOT module and a BLE module. By using the NB-IOT technology, the low-cost NB-IOT wireless intelligent lock has the characteristics that the power consumption is low; the signal coverage range is wide. The BLE module realizes the Bluetooth low energy consumption, so that the electric power is further saved; on one hand, the problem of short service life of the battery power supply equipment such as the intelligent lock is solved; on the other hand, the condition that a gateway needs to be additionally arranged for intelligent lock communication conversion is avoided. The device has the advantages of multifunction, low cost, energy conservation and environment protection.

The invention relates to an area information providing method and system based on a television terminal position and a real-time program. The method comprises a television terminal and a server side and comprises the following steps: the server side prestoring the position information acquired by the television terminal and monitors the video programs of multiple channels; the television terminal acquiring and uploading the video program information and simultaneously uploading the position information or television terminal information; the server side determining the position area and the video program that the television terminal belongs to; and the server side, according to the position area and the video program that the television terminal belongs to, providing the information to the television terminal. By using the television terminal information providing mode, area information can be provided to the television terminal according to both the position of the television terminal and also the video program broadcast by the television terminal; the area information is provided so that spectators pay attention to the information in a more convenient manner; and at the same time, by relying on the video programs, information can be provided according to the likings of the spectators.

The provided performance data report method for device single board comprises: (1) according to the normal fluctuation range of every performance index of board, setting a 'constant performance threshold' for every performance data; (2) collecting performance data; (3) comparing current monitored data with former threshold; and (4) if current data over the threshold, reporting the current data; or else, inhibiting the report. This invention reduces data quantity and CPU load, as well as the requirements of network element controller for data processing and storage room.

The invention discloses a combined putting system of advertizing service information, comprising an advertizing server, a service information module, a background processor, a combined distribution module and a service terminal interface, wherein the advertizing server is used for providing advertizing information and transmitting the advertizing information to the combined distribution module; the service information module is used for receiving the service request information from outside and transmitting the request information to the background processor; the background processor processes the service request information so as to form feedback service information which is stored and recorded and transmitted to the combined distribution module, and the service information module also selects the advertizing information type provided by the advertizing server according to the type of the service request information; and the combined distribution module puts the advertizing information and the feedback service information to the service terminal interface after combined distribution of interface display typesetting. When requesting service information, the system can pay attention to the advertizing information and improve the advertizing economic benefit.

The invention discloses a voice detection method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: inputting to-be-detected audio information into a voice classification model to obtain a first result, and enabling the first result to be used for determining a corresponding relation between theaudio information and a target object, wherein the voice classification model is obtained by training audio training information, and the voice classification model at least comprises a convolutionalneural network, a recurrent neural network and an attention mechanism model which are cascaded in sequence, and a fusion layer used for fusing output results of the convolutional neural network and the attention mechanism model. Thus, the audio information is detected based on the trained voice classification model to determine the corresponding relationship between the audio information and the target object, manual auditing is not needed, the labor cost and the time cost are reduced, the audio auditing efficiency is improved, and auditing and filtering of large-scale voice data can be realized.

The invention discloses a data processing method and a system based on a social platform and an e-commerce platform. The method comprises the following steps: acquiring marketing data of an account ofan e-commerce platform by establishing a connection with an interface system of the e-commerce platform; establishing an association between the e-commerce platform account and the social platform account; according to the obtained authorization, the marketing data of the e-commerce platform account is copied to the social platform corresponding to the social platform account, and displayed according to the setting interface. The proposal of the invention can effectively correlate the data of the e-commerce platform and the social platform, and the social platform can better attract the attention and participation of users.

The invention discloses a pedestrian re-recognition method combining posture and attention based on a double-flow network. The method comprises the following steps of: 1, preprocessing an input imageand inputting the input image into a double-flow network to extract features; and 2, combining middle-layer features and high-layer features and associating global information through an attention mechanism. And 3, fusing the attention flow and the attitude estimation flow through bilinear pooling operation to obtain a final feature map. And 4, model training: training neural network parameters byusing a back propagation algorithm. The invention provides a neural network model for pedestrian re-recognition, particularly provides a network structure combining an attention mechanism and attitude estimation based on a double-flow network, and obtains an effect of competitive power in the field of pedestrian re-recognition at present.

The invention provides a digital television program processing method, which comprises the following steps of: performing decoding processing on the data of a currently selected channel in a digital television signal to generate a program picture; determining a to-be-selected channel to be displayed on a television screen; performing the decoding processing on the data of the determined to-be-selected channel to generate the program picture; and displaying the program pictures of the currently selected channel and the to-be-selected channel by using playing windows with different attributes respectively. The invention correspondingly provides a digital television program processing system, which comprises a television signal processing module, a to-be-selected channel determination module, a to-be-selected channel image processing module and an image display module. By the method and the system, displayed images can be displayed in different ways in multi-picture display, and a user can visually pay attention to the program picture of the selected channel.

The invention discloses a recommendation-based interest community user guide method. The method comprises the following steps of S100, creating interest communities: mining potential users of all interests by analyzing an interest model of a recommendation system user, determining an order of the interest communities through a quantity of the potential users, and creating the interest communities in batches; and S200, recommending the interest communities: recommending the related interest communities to the proper users, and guiding the users to enter the interest communities. The method is suitable for quickly establishing interest communities of mobile internet recommended products and ensuring continuous importation of new users; and the users can obtain recommendation of the interest communities interested by themselves without search and can conveniently pay attention to the related interest communities, the browsing and internet-surfing efficiency of the users is improved, and various demands of the users are met.

The present invention provides a bus connector and a cabinet combination structure. The bus connector comprises an outer insulating sleeve, an inner conductive block is fixedly arranged in an inner cavity of the outer insulating sleeve, and conductive springs are arranged in a first plugging port and a second plugging port. The ends, away from the inner conductive block, of the conductive springsare electrically connected with a conductive plate, a color-changing drying agent is placed in the drying cavity, and a reset spring is connected with a first indicating rod. A connecting rope penetrates through a rope penetrating hole formed in a supporting piece, bypasses a guide wheel and then is fixedly connected to the conductive plate, a thermal expansion part is arranged on the heat absorption layer, and a second indication rod penetrates through a through hole formed in the side wall of the outer insulation sleeve. The bus connector provided by the present invention can know whether the desiccant is invalid, whether the internal elements are overheated and whether a bus connecting end is loosened or not on the premise of not opening the connector, and has the advantages of centralized functions and convenience in use. The present invention also provides a cabinet combination structure of the bus connector, and the cabinet combination structure is very worthy of popularization.

The invention belongs to the field of medical image processing, and particularly relates to a chest radiography multi-label classification method based on a self-correcting label generation network. According to the invention, a self-correcting label generation network model is constructed and is used for multi-label classification of chest radiography; the network comprises a ResNet image encoder, a self-correcting attention mechanism module and a GRU decoder, the image encoder is used for acquiring high-level semantic features of an input image, namely a chest radiography image, generating image feature representation and outputting the image feature representation to the self-correcting attention mechanism module; the self-correcting attention mechanism module generates an attention mapcorresponding to the current time step at each moment according to the state information and the image features output at the previous moment, and outputs a context feature vector to the decoder; andthe decoder generates a label corresponding to the image at the current moment according to the context feature vector and a label word vector generated at the previous moment. According to the method, the problem of unbalanced prediction effect in chest radiography multi-label classification is effectively solved.

The embodiment of the invention discloses an unsupervised image-to-image translation method. A specific embodiment of the method comprises the following steps: acquiring an initial image, and zooming the initial image to a specific size; performing spatial feature extraction on the initial image through an encoder to obtain feature information; inputting the feature information into a content style separation module to obtain content feature information and style feature information; generating reference style feature information of the reference image in response to the acquired reference image, and setting the reference style feature information as Gaussian noise consistent with the style feature information in shape in response to the fact that the reference image is not acquired; inputting the content feature information and the reference style feature information into a generator to obtain a target image translating the initial image into a reference image style; and zooming the target image to a size matched with the initial image to obtain a final target image. The implementation mode can be applied to various different advanced vision tasks, and the expandability of the whole system is improved.

The invention relates to a sitting posture recognition method based on projection reconstruction and a MIMO-CNN. The sitting posture recognition method comprises the following steps: acquiring a humanbody upper body depth image and a human body foreground contour map; performing pretreatment; projecting the depth information of the sitting posture contour, and reconstructing to obtain a three-view depth map; designing an MIMO-CNN network for sitting posture recognition and learning model parameters; sitting posture recognition; model self-learning. The method has the advantages that the preprocessed depth image is combined with the human body contour map, and interference of surrounding backgrounds on sitting posture recognition is eliminated. A three-view depth map is obtained by using aprojection reconstruction method, so that sitting posture information is richer. The invention discloses a designed MIMO-CNN structure. The method is especially suitable for projecting and reconstructing feature information and integrating an attention mechanism at the same time, hot spot areas of different sitting postures can be better concerned, so that the recognition precision is improved, meanwhile, model self-learning is adopted, the requirements for real-time performance and accuracy are well balanced, and the method has high anti-interference capacity for visual angle changes and complex environment backgrounds.

A method for sharing and tracking comments on a webpage comprises steps of reading a webpage via a first device reading a webpage, accessing at least one comment according to at least one piece of website information of the webpage from a second server or from at least one third device, and then displaying the webpage and the comment on the first device concurrently. Generating the comment includes steps of reading the webpage and generating the comment corresponding to the website information via a second device, and then applying the second server to save the comment and the website information. The present method allows users to share their comments corresponding to the webpage with each other and allows users to follow a specific webpage or a specific user's account and to manage an authorized user list.

The invention discloses a wave form brightness adjustment and correction method of a digital three-dimensional oscilloscope. According to hit probability of a signal sampling point, a user enables to-be-observed wave forms in a hit probability range to be displayed in a highlight mode through wave form brightness adjustment. In forward power law transformation, a wave form brightness grade increases with increase of the hit probability of the sampling point; and in reverse power law transformation, the wave form brightness grade reduces with increase of the hit probability of the sampling point. In the forward power law transformation and the reverse power law transformation, wave form brightness correction parameters can be adjusted to adjust wave form brightness distinguishability, and signals in different hit frequency ranges are conveniently observed. By the method of sectional transformation, the wave form sampling point falling in a certain range of the hit frequency can be specially focused on. According to the wave form brightness adjustment and correction method, wave forms in a specific frequency range can be observed and wave form display brightness can be changed at will, the defect of an existing digital three-dimensional oscilloscope is overcome, and the obtained wave forms can display more detail information.