41 results about "Event retrieval" patented technology

A computer system and associated methods for verifying user identities online. Identity claims made by a requestor of an online access and/or a trusted transaction may be verified by associating digital credentials to verified personal identification information (PII) retrieved from real world events. PII item(s) may be retrieved from third-party verified identity information sources. Verified personal attributes related to PII items may be identified and correlated with the requestor's digital credentials, and stored to a verified identity record. Additional digital credentials for the same requestor may be similarly identified, correlated, and stored to the verified identity record. A subsequent transaction request by a person claiming the requestor's identity may be compared with the verified identity record. An identity match indicator and/or a match confidence score may be created and used to determine the risk that the identity claim by the person requesting the transaction is false.

The invention relates to a traffic monitoring method and a traffic monitoring system. The method comprises the following steps: S1, collecting video information of a road through a camera; S2, performing multi-target detection and identification on the video information acquired in the step S1 to detect vehicles and pedestrians in the video; S3, performing tracking and trajectory analysis on the multiple targets detected and identified in the step S2; and S4, according to the multi-target tracking and trajectory analysis result in the step S3, obtaining the speed and acceleration of the target, judging whether a traffic collision accident occurs or not, and if the traffic collision accident is judged to occur, automatically carrying out alarm processing. Video collection is standardized, and video sources can be effectively managed; automatic analysis of video structuralization is established, and post-event retrieval depending on manpower is liberated; and compared with an existing traditional target detection algorithm, the traffic monitoring method and the traffic monitoring system are high in detection precision and good in effect; and multi-stream video structural analysis canbe input at the same time, and results can be output.

The present invention provides an improved hydraulic overshot tool. The overshot tool mechanically engages a cylindrical body downhole attempting to be retrieved. In the event retrieval is unsuccessful, the overshot tool may be hydraulically released from the cylindrical body. The release function employs a pressure differential within the overshot tool, but without necessity of a nozzle within the bore of the tool.

The invention relates to a fixed position based information prompting system and method. The system comprises a prompting board arranged at a fixed position and comprises a human body recognition module, a user identity determining module, a user event retrieval module and a prompting module. Preset user events include a user identity, events and/or articles. The human body recognition module is used for detecting whether a human body enters a recognition region. The user identity determining module is used for determining a current user identity. The user event retrieval module is used for obtaining the event and/or articles matched with the current user identity according to the current user identity and by retrieving the preset user events. The prompting module is used for prompting information of the prompting board fixing position according to the event and/or articles. The fixed position based information prompting system and method can automatically give a prompt to a user at proper time and at a proper place, the probability that a user forgets or misses things in life is remarkably reduced, and the user can conveniently and automatically obtain information associated with a specific position.

A method for determining relationship data associated with application programs in a messaging system, comprising the steps of: responsive to at least one first message event sending a message from a first application to a first destination and at least one second message event retrieving, by a second application, the message from a second destination, intercepting message data associated with the message; analysing the intercepted message data in accordance with one or more rules in order to find one or more message parameters; and in response to finding the one or more message parameters, identifying the first message event and identifying the second message event, determining a relationship associated with the first application and the second application. An apparatus and computer program element for determining such relationship data are also provided.

The invention discloses a news event displaying method and apparatus. The method includes the steps of obtaining retrieval information, inquiring a news event database according to the retrieval information, and obtaining candidate news event records including news event records matching with the retrieval information in the news event database where at least one news event record is preset. Each news event record includes news event occurrence site, title and keywords, wherein the keywords match with the retrieval information. The method includes the step of displaying the news event titles contained in the news event records at positions corresponding to the news event occurrence sites recorded in the candidate news event records on an electronic map. In this way, automatic calculation of geological information of news events is realized, and news event retrieval results are displayed according to different regions.

The invention provides an atmospheric environment tracing evaluation method and system based on an industrial park. The system comprises an abnormal event monitoring unit, an abnormal event retrieval unit, an abnormal event tracing unit, a pollution prediction unit, a management and control task distribution unit, a management and control task tracking unit, an event evaluation unit and a task flow form unit. The method comprises seven steps of testing, tracing, pre-managing, treating and evaluating, an atmospheric pollution tracing model, a pollution diffusion model and a task flow form are utilized to realize the whole-process closed-loop management of park atmospheric environment management work from early warning, timely response, accurate tracing and scientific treatment, the delay and blindness of conventional artificial offline judgment are reduced, and the management efficiency is improved. Early warning information receiving, pollution traceability and situation prediction of daily pollution events, proposing of enterprise governance measures afterwards, supervision and management of enterprise governance work and governance effect evaluation after rectification are achieved.

The invention relates to a specific sound event retrieval and positioning method based on sequence classification. According to the method, important context information is concerned by utilizing thetime sequence and attention mechanism of sound so as to extract sound deep features of a specific target sound event, and then a specific sound event retrieval network is trained by combining multi-task learning with regression loss and classification loss. When specific audio event retrieval and positioning are carried out on a given audio file; the method comprises the following steps: firstly,inputting Mel characteristic energy of a sound segment to be detected into a sound retrieval model to obtain a retrieval result of a specific sound event of each sound segment, then positioning startand stop audio frames appearing in the specific sound event through post-processing, and finally obtaining complete specific sound event retrieval and positioning information of an audio file throughsmoothing processing.

The multi-level vision-text semantic alignment model MSAVT used for image-text matching is provided, the news event retrieval method based on the multi-level vision-text semantic alignment model MSAVT used for image-text matching is provided, news event cross-modal image-text search is achieved, and the current news retrieval requirement is met. The image-text alignment precision of the cross-modal retrieval model provided by the invention is higher, and when the cross-modal retrieval model is applied to news event cross-modal image-text retrieval, indexes such as recall rates of multiple levels, average accuracy and the like are remarkably improved. And meanwhile, a pre-trained BERT model is introduced to extract text features, so that the generalization performance of the algorithm is improved. The model adopts a public space feature learning method, vector representations of images and texts can be independently obtained, namely, vector representations of retrieval results can be stored in advance, retrieval time is short, and the method can be applied to actual scenes.

The present invention discloses an energy control emergency preplan processing method based on graphic modeling, comprising the following steps: compiling an operating plan by using graph programming, and describing the operating plan in a form of plan flow chart; storing the operating preplan to a preplan library, and establishing a correspondence between the operating preplan and a meta-event in a preplan library event pool; monitoring the event pool; when an event occurs, querying the event pool; and listing, by means of event retrieval and adaptation, a possible operating preplan that meets a condition, to provide operating personnel with a corresponding preplan flow chart. After determining that the operating preplan is inerrant, the operating personnel select to automatically execute the operating preplan or manually execute the operating preplan in a single step. According to the energy control emergency preplan processing method based on graphic modeling, experience of enterprise fault emergency treatment is compiled to a preplan in a form of plan flow chart by using graph programming, and the preplan is automatically executed by using an automation control means when a fault occurs or a plan triggering condition is met, thereby achieving an objective of fast executing an operation and accurately recovering a fault.

Time-varying scent determined by audio-meaningful events, such as gun firing, in an audio signal is generated by first computing a spectrogram thereof. A fingerprint, each having feature points, can be extracted from the spectrogram over a time window. The plurality of feature points is hashed to give a hash code. A sequence of hash codes obtained over multiple time windows is correlated with predetermined hash-code segments of known audio-meaningful events stored in a database. A matched audio-meaningful event having a highest correlation is identified. The scent-generating recipe corresponding to the matched audio-meaningful event is retrieved. Atomizers are used to vaporize scent-generating substances to generate the time-variant scent. Since the hash code has a fixed length and a smaller size than the plurality of feature points, storage requirements of the database and computation requirements of correlation calculation are reduced by using the hash code rather than the plurality of feature points.

The invention provides a retrieval method of events and companions. The method is suitable for use in the field of internet technology. According to the embodiment of the invention, finding is carriedout according to an input query user and query keyword, and results including the pairing people of the query user and the matching events of the query keyword are obtained. The embodiment of the invention fuses the advantages of event and companion recommendation and keyword queries, retrieves event and companion pairs, and comprehensively considers text relevance related to text files and peerrelevance of expectation degrees of the query user on the peer companions. According to the embodiment of the invention, both text similarity and relationships of the events and the people are considered in event retrieval, a query can be carried out according to the query user and the query keyword, running time of the single query is at a second level, and query speed is high.

The invention relates to an intelligent mass surveillance video analysis system. An analysis method comprises the following steps: S1, inputting the time, place and mark of an event into a retrieval system, S2, dispatching the video of the event input into the retrieval system, S3, conducting video enhancement and night video processing on the video parameters of the dispatched video or concentrating abstracts and analyzing to obtain clues or obtaining details of a case through manual inspection or conduction quick search according to classifications of video targets, S4, inputting the clues and event description output from S1 to S3 into an image inspection and recording subsystem for future searching. By means of the intelligent mass surveillance video analysis system, a deblurring technology which is developed aiming at movement blurring and based on blurring core evaluation is achieved, hardware space for storing historical videos is saved by 90%, possible locations of the targets are found quickly and a quick list is produced for users to search directly. Further, the intelligent mass surveillance video analysis method is environmental friendly.

The invention provides an event reminding method and terminal. The event reminding method comprises: a first event list is retrieved regularly at a terminal and evens which are not dealt with are generated into a second event list; according to a preset classification rule, the events in the second event list are classified; and according to a classification result, a regular retrieving times number, and a set reminding way, event reminding is carried out. Besides, the event reminding terminal consists of a central processor, a communication module, an event retrieval module, an event classification module, an event reminding module, a selection setting module and a battery module; and the central processor is connected with the communication module, the event retrieval module, the event classification module, the event reminding module, the selection setting module and the battery module. Therefore, reminding of commercial tenant event dealing with is promoted; and missing processing is avoided.

The embodiment of the invention provides a method and a system for obtaining answers to food safety questions based on LSTM. The method comprises the steps of obtaining a to-be-answered question; inputting the to-be-answered question into a pre-trained question and answer judgment model, and obtaining a question matching result output by the question and answer judgment model; wherein the questionand answer judgment model is obtained by training an LSTM model by using manually labeled question and answer pairs and corresponding matching labels based on a food safety question and answer sampleset. According to the method of the invention, the food safety questions are trained based on the LSTM model, the problem of low food safety event retrieval accuracy of a search engine is solved, andanswers to the food safety questions can be effectively screened out from a news library.

The invention discloses a method for realizing extrasensory experience by retrieving video content. The method comprises three steps of video preprocessing, video event retrieval and video event retrieval result processing, and relates to the technical field of video event retrieval and extrasensory device mapping. The method disclosed by the invention aims to realize automation of introduction of four dimensional stunt in a 4D (four dimensional) cinema, compared with the method of manual introduction of the 4D stunt in the 4D cinema in the prior art, the method disclosed by the invention realizes the automation of the introduction of the 4D stunt at the greatest extent, the efficiency of the introduction of the 4D stunt in the 4D cinema is improved at the great extent, and the more real 4D experience can be provided for the audience.

In accordance with one embodiment, an information processing apparatus, which is connected with a server, comprises a printer configured to form an image on a recording medium, an input device configured to input information, a log file to be uploaded to the server, a setting file which stores, for each log file, presence/absence of execution of uploading for each event, and a control section configured to acquire, if it is determined that the event has occurred, the category of the event that has occurred, retrieve the setting file with the acquired event, read the name of the log file for which presence of execution of the uploading in the event is detected, and then upload the log file specified with the log name to the server.

The invention discloses an event retrieval method and device, computer equipment and a storage medium. The method can comprise the following steps: training to obtain an identification model for identifying whether two event descriptions correspond to the same event or not; forming an event vector extraction service based on the identification model, wherein the event vector represents understanding of the identification model for event description; according to the event vector extraction service, respectively obtaining event vectors corresponding to event descriptions to be retrieved and event vectors corresponding to event descriptions of events in an event library; and according to the obtained event vector, determining an event matched with the to-be-retrieved event description in anevent library. By applying the scheme of the invention, event retrieval based on event description can be realized, and the accuracy of a retrieval result is improved.

The invention relates to a rapid seismic event retrieval method and device, and has the beneficial effects that a set of quick retrieval method is designed aiming at the influence characteristics of earthquakes, and the method can quickly retrieve seismic events which may cause loss. The method of the invention first needs to create a spatial index of a seismic event. Although the traditional index can quickly retrieve the position of data, as related data are often stored in different places, the whole block cannot be read at one time, and the data retrieval speed is reduced due to frequent reading operation. The storage sequence and position of the data are reorganized according to the spatial position of the earthquake, and the time complexity of retrieval is ensured to be O (1).