62results about How to "Flexible identification" patented technology

A wearable device is provided that is used while being worn on a body of a user. The wearable device includes a biometric-information authentication portion that acquires biometric information from the user for biometric authentication. The wearable device acquires biometric information from a portion in contact with or in the vicinity of the user's body to perform the biometric authentication without requesting the user to operate for the biometric authentication.

The invention discloses a piece of automatic walking equipment and a working region judgment method thereof. The method comprises the following steps: an image of a target region is shot; the image is divided into a plurality of sub image blocks; colors of pixels of at least one sub image block are extracted; the proportion of a predetermined color in the sub image block is calculated and compared with a first preset value; the texture feature value of the sub image block is extracted and compared with a second preset value; and a sub region corresponding to the sub image block is judged as the working region when the proportion of the predetermined color in one sub image block of the image reaches or exceeds the first preset value and the texture feature value reaches or exceeds the second preset value. Therefore, the working system is enabled to be simple and user-friendly, and working region identification is flexible and convenient.

The invention provides a mobile payment security protection method, a mobile payment security protection device and a cloud server. The method comprises the following steps that: when a trigger event of a mobile terminal payment type client is monitored, the feature information of the payment type client is obtained, wherein the feature information refers to a unique identify label of the payment type client; the obtained feature information is matched with the feature information preset in a white list, and a matching result is obtained, wherein the white list comprises the feature information of the mobile terminal payment type client capable of securely completing payment type operation; whether the payment type client is a malicious client or not is judged according to the matching result; and the payment type client is processed according to the judging result. The technical scheme provided by the invention has the advantages that the security of the payment type client on a mobile terminal can be improved, and the protection performance on hostile attack of viruses, Trojan horses and the like is enhanced, so that a secure and clean mobile pavement environment is provided for users.

For recognizing traffic signs that include a main sign and an associated additional sign, a method involves using a camera to record image data of an environment outside a vehicle, evaluating the image data to detect the presence and the class of the main sign and to detect the presence and partially classify the content of the additional sign by pattern recognition, performing text recognition to read and interpret text of the additional sign, and comparing the recognized text with actual current situational information to determine whether the main sign is currently applicable.

The invention discloses a semantic recognition method and device, and relates to the technical field of computers. The method comprises the specific steps that to-be-recognized statement information is acquired; the to-be-recognized statement information and preset statement templates are matched based on the longest common subsequence, and a matching result with a weight value is determined; semantic recognition is carried out according to the matching result. The method can perform more accurate semantic recognition on the acquired to-be-recognized statements on the basis of the matching method with the weight value of the longest common subsequence according to the preset statement templates, completely utilize statement template information, and be higher in flexibility and high in efficiency. In addition, the method can update the statement templates in real time in a short time to perform subsequent testing feedback.

The invention discloses a detection method for a curved form line of a financial statement. The detection method comprises the steps of obtaining a form image and performing the image preprocessing; obtaining the actual character height and performing scaling processing; removing the red seal; carrying out the convolution processing to enhance horizontal lines and vertical lines of the table; carrying out the adaptive binarization processing to obtain a binarized image; filtering out non-form line interference; carrying out completion and interference elimination processing on the horizontal line and the vertical line by adopting a creeping worm algorithm, adding the horizontal line graph and the vertical line graph to obtain a table image, correcting the table image according to table intersection characteristics, correcting a bent table line and recognizing characters, and reconstructing a table. According to the method, the table is reconstructed, and the key information in the table is extracted, so that the table line positioning problem of the curved table is solved, and the mapping, storage and verification from the financial statement to the standard table are met.

The invention provides a transaction request processing method and device and a distributed system. The transaction request processing method comprises the steps that a transaction request is received; a request parameter in the transaction request is extracted; the request parameter is matched according to a filter condition in a determination model, wherein the determination model comprises a number of filter conditions generated by the combination of exception parameters extracted from exception processing logs; and if the request parameter matches the filter condition, the transaction request is rejected. According to the transaction request processing method provided by the invention, the recognition accuracy of an exception request is improved; the exception request can be effectively and efficiently identified; and the determination model is not easily cracked.

The embodiment of the invention provides an abnormal domain name identification method. Domain name to be identified can be acquired, text features of the domain name to be identified are extracted; the text features of the domain name to be identified are input into a machine learning model; an identification result of the to-be-identified domain name output by the machine learning model is obtained; wherein the machine learning model is a classification model trained based on a black and white sample set; wherein the black-white sample set comprises black samples and white samples; wherein the black sample is a known abnormal domain name; wherein the white sample is a known normal domain name; Compared with the prior art in which the domain name is identified by means of a known abnormaldomain name or a known abnormal domain name forming rule, the method has the advantages that the accuracy is higher, and the domain name is identified by means of a pre-constructed machine learning model, so that the domain name identification is more flexible.

The embodiment of the invention provides a table recognition method and device, a terminal and a storage medium. The method comprises the steps: acquiring a to-be-recognized file which comprises at least one table; utilizing a table position detection model to detect position information of at least one table in the file, the table position detection model being a deep learning model; processing the file according to the position information to obtain a table picture which corresponds to each table and only contains the table; carrying out the following processing on each table picture: performing pre-processing operation, processing the processed table picture according to a table row division model and a table column division model, and acquiring row position information and column position information of the table, wherein the table row division model and the table column division model are both deep learning models, so that the structure information of the table is restored; and outputting structured table data in combination with the recognized text information of the table. In the way, the table recognition efficiency and flexibility can be improved.

The embodiment of the invention discloses an image processing method and electronic equipment, relates to the technical field of communication, and aims to solve the problem of how to enrich modes ofidentifying the identity of the electronic equipment. The method is applied to first equipment, and comprises the steps of obtaining a target image used for indicating configuration information of second equipment; identifying a first graph in the target image; determining a first character corresponding to the first graph according to the main element of the first graph and the auxiliary elementof the first graph; identifying a second equipment based on the first character. The method can be applied to a scene of equipment identification.

The invention relates to a control method and a control device for a home appliance. The method comprises the following steps: acquiring user information; matching the user information in a preset user information library; when the user information is successfully matched in the preset user information library, acquiring setting information corresponding to the user information; setting the home appliance according to the setting information. The method and the device have the beneficial effects that the acquired user information is matched with preset user information in the preset user information library, so that different users can make different settings to the home appliance, and for example, the addition of a detergent, a washing mode, washing time and the like of a washing machine can be set for different users. Therefore, a novel home appliance control method is provided for the user, and meanwhile, good intelligent experiences are brought to the user.

The invention belongs to the field of 3D printing, particularly relates to a 3D printing online quality monitoring method, system and device based on a neural network, and aims to solve the problems that an existing defect detection method can only detect specific defect shapes and is low in defect detection precision. The method comprises the steps of collecting an image of a three-dimensional object in the 3D printing process and using as an input image; acquiring the category of each pixel in the input image by adopting a pre-trained defect segmentation network; counting the pixel number corresponding to each type of defect in the input image, and calculating the area of the defect part in the input image by combining with the pre-acquired camera internal reference; and judging whether the area is larger than a set threshold value, if yes, starting a quality monitoring alarm, and if not, images in the 3D printing process continue to be collected. According to the method, 3D printing defects of different types and shapes can be flexibly identified, the false detection rate is reduced, and the defect detection accuracy is improved.

The invention relates to the technical field of automation, in particular to a processing line mistake proofing system and method, a storage medium, electronic equipment and a production line, and solves the problems that automatic production line detection equipment is complex and inconvenient to install and maintain, and the production line cannot be stopped in time and warned due to long detection time. The system is characterized in that a sensor respectively sends in-place signals to a roller way control device and an upper computer when detecting a to-be-detected workpiece; an image acquisition device acquires an image of the processing characteristics of the to-be-detected workpiece; an upper computer is respectively connected with the image acquisition device and the roller way control device, and is used for controlling the image acquisition device to acquire an image based on the in-place signals, processing the image, judging whether the image contains to-be-detected features or not, judging whether the to-be-detected features meet requirements or not, and feeding back preset information corresponding to a judgment result according to the judgment result; and the roller way control device is connected with the upper computer, the sensor and the roller way and controls running and stopping of the roller way according to preset information.

The invention relates to a flexible production line based on visual inspection. The flexible production line comprises an upper computer, a control module, a visual inspection device connected with the control module and a workpiece conveying mechanism for conveying a to-be-inspected workpiece; the visual inspection device is erected on the workpiece conveying mechanism and inspects the to-be-inspected workpiece to identify defects; after the to-be-inspected workpiece is identified to have the defects, the control module controls a mechanical pick-up arm to pick up the to-be-inspected workpiece and transfers the to-be-inspected workpiece to another workpiece conveying mechanism; and the control module reports an inspection result to the upper computer. With the flexible production line, the workpiece inspection efficiency is improved by the aid of the visual inspection device, multi-angle images of the to-be-inspected workpiece can be acquired, the workpiece is inspected more comprehensively and flexibly, the defective workpiece is picked up automatically after being identified, and the production efficiency is improved.

The invention relates to a webpage text editing method and device based on URL recognition and a storage medium, and belongs to the technical field of the Internet. The webpage text editing method comprises the following steps: reading a URL link in a webpage text; identifying the type of the URL link, and calling a corresponding template; and displaying the content corresponding to the URL link according to a template format. The webpage text editing device comprises a processor and a memorizer, wherein a computer program is stored in the memorizer, and when the computer program is executed by the processor, the webpage text editing method based on URL recognition is achieved. A computer program is stored on the medium, and when the computer program is executed by the processor, the webpage text editing method based on URL recognition is achieved. Compared with the prior art, the webpage text editing method has the advantages of being convenient to edit, high in automation performance, wide in application range and the like.

The invention relates to the technical field of washing machines and aims at solving the problem that existing washing machines cannot flexibly identify the materials and color of washed objects and cannot prevent color crossing of the washed objects. In order to solve the problem, the invention provides an intelligent detection module, a washing machine and a washing procedure recommendation method of the washing machine. The intelligent detection module is detachably connected with a washing machine body. The intelligent detection module includes an identification module which can identify the materials and / or color of the washed objects in the disassembly state and / or fixed state of the intelligent detection module. The washing machine includes the intelligent detection module. The recommendation method includes the following steps that the intelligent detection module is close to the washed objects or the washed objects are close to the intelligent detection module so as to identify the materials and / or color of the washed objects, and according to the identified material and / or color information, a washing procedure is recommended. The intelligent detection module can flexiblyidentify the materials and color of the washed objects, and prevent color crossing of the washed objects.

The present invention provides a container number identification method and device for a carrying vehicle and a carrying vehicle. The method comprises: detecting whether a container and a carrying vehicle has the relative movement or not, after it is detected that the container and the carrying vehicle have the relative movement, collecting the image information of the container numbers at the at least two moments according to a preset collection mode; and processing the image information of the container numbers according to the preset processing mode to generate container number information. According to the technical scheme, the container number identification method and device for the carrying vehicle and the carrying vehicle can process the container number identification with a plurality of printing modes to realize the container number identification function of the container at different operation states, improve the efficiency of the container number identification and satisfy the user usage requirement.

The invention relates to the technical field of washing machines, and an objective is to solve problems that a fluorescent agent and / or pigments cannot be detected flexibly and conveniently, and accumulation of the fluorescent agent and / or pigments in a washing machine cannot be prevented. For this purpose, the invention provides a pollutant monitoring module used for a washing machine, a washingmachine, and a control method thereof. The pollutant monitoring module can be connected with a washing machine body in a detachable manner, and the pollutant monitoring module comprises an identification module. In a dismounted state and / or a connected state of the pollutant monitoring module, the identification module can identify pollutants in a detected object. The washing machine comprises thepollutant monitoring module. The control method comprises: making the pollutant monitoring module close to clothes or making clothes close to the pollutant monitoring module, to identify whether clothes pigments and / or the clothes have fluorescent agents; and according to information of the identified pigments and / or the fluorescent agents, recommending a washing program. The module can flexiblydetect the fluorescent agents and / or the pigments, and absorb the fluorescent agents and / or the pigments.

The invention relates to a flexible production line, comprising a control module, and an automatic discharging and feeding mechanism, a vision detector, a mechanical picking arm and a workpiece transferring mechanism for transferring workpieces to be detected, which are connected with the control module. The automatic discharging and feeding mechanism is used for conveying the workpieces to be detected to the workpiece transferring mechanism; the vision detector is arranged on the workpiece transferring mechanism in an erecting way, and used for detecting the workpieces to be detected; and when the defect of a workpiece to be detected is recognized, the control module is used for controlling the mechanical picking arm to pick the workpiece to be detected. According to the flexible production line provided by the invention, the workpiece detecting efficiency is improved through the vision detector; a multi-angle image of the workpiece to be detected can be acquired, so that the detection can be carried out more comprehensively and flexibly; and the defective workpiece is automatically picked after being recognized, so that the production efficiency is improved.

The invention discloses a live communication terminal capable of automatically carrying out bidirectional communication after a unidirectional call, a method and a tool installed on a mobile terminal, wherein the live communication terminal capable of automatically carrying out bidirectional communication after the unidirectional call receives a connection request from a credible user; the received connection request from the credible user is responded, and an answer for the connection request is automatically sent, so that IP (Internet Protocol) communication with the credible user is automatically established; the collected video frequency and voice frequency are transmitted to the credible user in the IP communication with the credible user, and the voice frequency from the credible user is at least received. In comparison with the prior art, the connection request of the credible user is automatically responded through the live communication terminal capable of automatically carrying out bidirectional communication after the unidirectional call, so that the interaction of the credible user at a monitoring end and a monitored end is enhanced, thus the communication experience of the user is improved.

The invention provides a face recognition method and a face recognition system for detecting a shielded area in the face. The face recognition method comprises the steps of performing face detection on an image, and marking the detected face; performing positioning on the marked face for acquiring a central point of a left eye and a central point of a right eye; forming a geometrical area according to the central point of the left eye, the central point of the right eye and two top points at upper side of marking frame; calculating the chroma component of each pixel in the geometrical area; screening different pixels of which the chroma components are different from a preset chroma, performing statistics on an interval which corresponds with the different pixels and using the interval as a shielded area. According to the face recognition method and the face recognition system, through performing statistics on the pixels of which the chroma components are different from that of a skin color in the geographical area that is composed of a marking frame and double eyes and using the obtained interval as the shielded area, the face recognition method and the face recognition system have advantages of no requirement for sample and training parameter collection, no interference by user behaviors, high accuracy, high calculation speed and high recognition flexibility.

The invention relates to a mobile face recognition method and system, a computer and a storage medium. The method comprises the following steps: shooting by wearable image shooting equipment to obtaina moving image; analyzing the moving image to obtain a face image in the moving image; performing face comparison on the face image and a pre-stored person image in a local storage device, and detecting whether the pre-stored person image matched with the face image exists in the local storage device or not; and when it is detected that the pre-stored person image matched with the face image exists in the local storage device, sending an alarm signal. According to the invention, the portable wearable image shooting equipment shoots and obtains the moving image, analyzes the moving image to obtain the human face image, compares the human face image with the personnel image pre-stored in the local storage device, detects whether the personnel in the moving image is the key personnel, and can shoot the suspected key personnel more conveniently and flexibly from multiple angles, so the contrast efficiency is effectively improved.

The invention relates to a submerged marine organism automatic identification and sampling system. The system comprises two buoys connected through a connecting rod, and detection rods arranged at thebottoms of the two buoys; a receiver is arranged on the outer side of the lower end of one detection rod, the bottom of the other detection rod is connected with a light source face matched with thereceiver, a plurality of lamp strips are arranged on the light source face in a staggered mode and can emit light curtains with different wavelengths, and the light curtains emitted by the lamp stripsare overlapped to form a light field. Grooves are formed in the buoys in the circumferential direction, clamping blocks are arranged at the two ends of the connecting rod respectively, the clamping blocks at the two ends of the connecting rod are connected with the grooves in the two buoys in a matched mode respectively, and the connecting rod can rotate in the circumferential direction of the buoys. The receiver can recognize marine organisms through structured light curtain light source signals, and a sampling piece is arranged on the connecting rod and can collect the marine organisms.

The invention provides a data flow identification method and device. The method is applied to network equipment, and comprises the following steps: receiving a data packet sent by intrusion preventionequipment; finding a corresponding data packet feature table item from a preset data packet feature table according to the packet feature information of the data packet; if the abnormal level of thedata stream is lower than the specified level, inputting the data packet into a data stream identification model matched with the abnormal level of the data stream, and identifying whether the data stream to which the data packet belongs is an abnormal data stream or not according to an output result of the data stream identification model; and if the abnormal level of the data stream is higher than the specified level, identifying the data stream to which the data packet belongs as an abnormal data stream. On the basis of the intrusion prevention equipment, supplementary identification is carried out on the data stream identified as the non-abnormal data stream by the intrusion prevention equipment, so that the defect that the identification range of the intrusion prevention equipment onthe abnormal data stream is relatively small is overcome.

The invention discloses a license plate recognition method based on sequence learning, and belongs to the technical field of intelligent traffic. A license plate image is changed into a sequence composed of candidate characters through a deep neural network used for character detection and recognition, then the sequence is input into a sequence conversion model obtained through training, a new character index sequence is obtained, and finally original candidate characters are recombined according to indexes to obtain a license plate recognition result. According to the license plate recognition method based on sequence learning obtained through the technology, a machine autonomously learns license plate character selection, rejection and permutation combination rules, the pressure of manually designing heuristic rules is reduced, learning experience is achieved with the help of a large number of data samples, and a sequence model can be more flexible in the aspect of the recognition problem of multi-standard license plates. Compared with a traditional sequence processing method, such as hidden Markov, the method has the advantages that the connection of the sequence elements can be found in long-time dependence, and a better sequence conversion result is obtained.

The invention discloses an intelligent dialogue system and method based on robot deep learning and a storage medium, and belongs to the technical field of artificial intelligence. Comprising the following steps: acquiring voice data for voice recognition to obtain a transliteration text of the voice data; constructing an information tree based on the transliteration text, and creating sentence segmentation nodes according to Chinese grammar; segmenting the transliteration text based on the sentence segmentation nodes and obtaining multiple groups of parallel short sentences; questioning voices are obtained, subjects in the questioning voices and corresponding questioning types are recognized, and a retrieval sequence is generated; and inputting the retrieval sequence into a matching model, traversing the sub-branches, the branches and the information tree by the matching model to find an optimal answer, and broadcasting the optimal answer. The method and the system have the learning and scene deep session capability, and besides a question-answer language processing method, the method and the system can deeply memorize and understand the meanings of the questions and retrieve the answers of the questions.

The invention belongs to the technical field of battery management, and particularly relates to an analog front-end acquisition system based on battery management, which comprises a multi-channel data acquisition module, an analog front-end unit, an FPGA chip, a storage module, an intelligent label generation module, an intelligent label card reading module, an upper computer, a data abnormity monitoring unit, an alarm module and a wireless communication module. The output end of the multi-channel data acquisition module is connected with the analog front-end unit, and the multi-channel data acquisition module is used for battery data acquisition. The analog front-end unit is used for carrying out filtering, amplification and analog-to-digital conversion processing on received sensing signals, the carrier-to-noise ratio of transmission signals is increased, stable and safe data transmission is achieved between the FPGA chip and the upper computer through the wireless radio frequency identification technology, mechanical or optical contact between an identification system and a specific target does not need to be established, penetration and barrier-free reading can be achieved, the intelligent label can be repeatedly used, updating of information data is facilitated, and identification is accurate and flexible.

The invention belongs to the technical field of network security and machine learning, and provides an attack organization identification method based on industrial control traffic and threat intelligence association analysis, which comprises the following steps of: obtaining traffic data, threat intelligence information, targeted characteristics and a scanning tool of an industrial control attacker; aiming at the obtained data characteristics of the threat intelligence characteristics and the industrial control traffic characteristics of the attacker, respectively adopting different similarity measurement methods to construct a similarity matrix, fusing to construct a composite weighted similarity matrix, and mapping data points in an original high-dimensional space to a low-dimensional space; clustering the obtained data sample points to obtain an industrial control attack organization; according to the method, the recognition precision of the attack organization is improved, the expandability is high, and the method can be used for monitoring network intruders in real time and actively defending distributed attacks of the attack organization in time. The number of clusters does not need to be specified, the bandwidth can be adaptively calculated, and industrial control attack organizations can be identified more accurately and flexibly.