35results about How to "Improve recognizability" patented technology

The invention discloses an image identification method of stirring friction welding arc stripe interval. The image identification method is characterized by mainly comprising the steps of: obtaining image, carrying out binarization processing on the image, carrying out post processing on the binary image and detecting and calculating the arc stripe interval, firstly, obtaining a microscopic picture of a stirring friction welding arc stripe tool; and then, performing gray scale, wave filtering and binarization processing on the microscopic picture of the arc stripe tool; removing the isolated points, the interior stripe detects and the incomplete edge stripes of the binary image; and finally, determining the arc stripe interval calculation range, calculating the arc stripe interval and processing the data. The image identification method of stirring friction welding arc stripe distance disclosed by the invention greatly enriches the obtaining method of arc stripe parameters and provides certain foundation for research of the arc stripe parameters in the stirring friction welding.

The invention discloses a brain tumor image automatic segmentation method based on a convolution neural network, which comprises the following steps: acquiring a multimodal MRI image of the brain tumor and carrying out image preprocessing to obtain an original image set; constructing the frame of brain tumor segmentation based on multimodal MRI images; the frame comprises a module 1 and a module 2. The module 1 comprises a parallel deep deconvolution neural network based on a 3d convolution neural network, a residual unit and a transposed convolution; the module 2 includes adding a hopping structure on the basis of the deep deconvolution neural network structure in the module 1; after several iterative training, the ideal weights are obtained and the segmentation images of brain tumors areoutput. Image segmentation results are tested and evaluated. The invention effectively solves the problem of low accuracy of brain tumor segmentation image segmentation. It can improve the recognition of tumors and make the image preprocessing more convenient. Using the loss function based on d coefficients in the segmentation module can effectively solve the problem of class imbalance.

The invention provides a method and a device for generating a photo album cover. The method comprises the steps of acquiring at least one picture included by a photo album, and identifying image features included by each picture; selecting one or more pictures of which the picture features meet preset feature information from the at least one picture according to the identified picture features included by each picture; and generating the cover of the photo album by using the selected one or more pictures. According to the method provided by the embodiment of the invention, by identifying the picture features of each picture, and selecting the picture of which the picture feature meets the feature information preset by a user as the cover of the photo album, the cover of the photo album can more effectively and accurately reflect more information of the photo album, and thus the identifiability of the content of the photo album is improved.

This invention belongs to bank bill codes network identification test monitor inquire anti-false system and method, wherein, the anti-false system comprises wireless mobile network, bank process message invisible codes print monitor and inquire anti-false system and user holder mobile scanning identification code terminal system. The invention method comprises the following steps: establishing the bank bill invisible code print monitor system and inputting personal or business information with check business into data process host and to update database.

The embodiment of the invention discloses a method and device for extracting image features based on semi-supervised learning. The method comprises steps of: initializing a model parameter and preprocessing image data to obtain image samples, and dividing the image samples into training samples including tagged samples and untagged samples, and test sample just including the untagged samples; according to a pairing constraint condition, determining a constraint set corresponding to the tagged samples in the training samples; by using a neighbor search algorithm, constructing neighbor graphs corresponding to all the training samples and calculating a weight matrix; by minimizing a feature approximate error, subjecting the training samples to low-dimensional manifold feature processing to obtain low-dimensional manifold features and a linear projection matrix; and extracting the image features of the training samples and the test samples by using the linear projection matrix. The methodand device can simultaneously maintain the global and local structure information of the sample data, improve the distinguishability of the features, quickly map new test data to low dimension and improve the performance of image feature extraction.

The invention relates to a design method for linear discrimination of a sparse representation classifier based on nuclear space. The method comprises the following steps of reading training samples, performing nonlinear transformation on the training samples to transform the training samples to the high-dimensional nuclear space, learning the training samples of each kind in the high-dimensional nuclear space, finding out the contribution (namely the weight) made by each individual in the training samples of the kind to constructing the subspace of the training sample of the kind, forming dictionaries through products of the training samples of the kind and a weight matrix, and sequentially arranging the dictionaries of all kinds to form a large dictionary matrix; obtaining linear discrimination sparse codes of the test samples inside the nuclear space on the basis of the dictionary matrix, and performing fitting on the test samples through the dictionaries of each kind and linear discrimination coding corresponding to the dictionaries; adopting the kind with the minimum fitting error as the category of the test samples. It can be ensured that sparse codes of the samples of the same kind are concentrated, sparse codes of the samples of different kinds are dispersed, the sample discrimination is effectively improved, and the performance of the classifier is improved.

The present invention discloses a strange number processing method and device: when a call from any calling number is received and the calling number is determined to be a strange number, determining whether the calling number is stored in a strange number marking database; displaying the calling number and marking information corresponding to the calling number stored in the strange number marking database to the user if so; otherwise, after connecting this call and ending a conversation, guiding the user to create marking information corresponding to the calling number, and correspondingly storing the marking information created by the user and the calling number into the strange number marking database. The present invention discloses a strange number processing device at the same time. With the solution of the present invention, the identifiability of the strange number is improved.

A head light device of vehicle capable of inhibiting enlargement of upper and lower sizes of the head light as well as improving probability of position lamp being identified is provided. The head light and the position lamp are accommodated a head light unit (30) of a single housing (31) and the position lamp bulb (60) is configured in a position on a lower end of a reflector (43) of the head light bulb (40) and in displacement in any one of left and right sides of the width direction. There are a pair of left and right position lamp bulbs (60) in left and right sides in width direction on lower end of the housing (31). A pair of left and right indication lamp bulbs (50) are arranged in left and right sides in width direction on upper end of the head light bulb (40). A pair of left and right head light bulbs (40) is arranged in a gap that is narrower than that of the position lamp bulbs (60). A light guide plate (41) of the head light bulb (40) and a reflector (61) for the position lamp are integrally formed to reduce parts number.

A display device includes a display panel and a control unit. The control unit is configured to control a display condition of the display panel by successively switching a plurality of frames, thereby to display a pointer image indicating a pointer and a path image indicating a path of motion of the pointer image. The path image of a first frame, which is any one of the plurality of frames, is provided continuously from a second location where the pointer image of a second frame, which is another one of the multiple frames and immediately before the first frame, is displayed to a first location where the pointer image of the first frame is displayed.

The invention discloses a fuzzy license plate reconstruction method based on characteristic learning. The method comprises the following steps of acquiring a large number of clear character samples covering all license plate character types and acquiring corresponding fuzzy character samples; setting up a fuzzy character sample database and performing characteristic training; acquiring fuzzy license plate images and calibrating four angular points of last five characters of the images; performing geometric correction and character segmentation to acquire five segment characters; acquiring the clear character sample and the fuzzy character sample corresponding to each segment character; according to the clear character sample and the fuzzy character sample corresponding to each segment character , performing character reconstruction on the segment character and then acquiring a reconstruction result; and performing histogram processing on the reconstruction result. The reconstruction result identifiability can be greatly improved. Relatively good reconstruction effects are achieved. License plate images can be recovered in a relatively clear and accurate way. The method can be widely applied in the license plate number identification field.

A structure of vehicular illuminator can ensure enough brightness of bulb light transmitting to different directions and can execute reliable light distribution in the structure of vehicular illuminator which executes light distribution in the direction different from the reflection direction of main reflector. The structure of vehicular illuminator is provided with the following components: a notch (81) which is formed on a main reflector (67) and guides the bulb light to the back side of main reflector (67), and an auxiliary reflector (82) which extends to the back part of main reflector (67) along the irradiation direction of bulb light guided by the notch (81). The bulb light which passes through the notch (81) and is guided to the back part of main reflector (67) is executed with light distribution in a direction different from the reflection direction of main reflector (67) by the auxiliary reflector (82).

The invention relates to a feature coding model generation method, an audio determination method and a related device. The feature coding model generation method comprises the following steps: acquiring a plurality of sample audios marked with category labels; extracting audio features of the plurality of sample audios; encoding the audio features of the plurality of sample audios through a feature encoding model to obtain a plurality of encoding vectors of the plurality of sample audios, and classifying the plurality of sample audios according to the plurality of encoding vectors to obtain category predicted values of the plurality of sample audios; and determining a target loss value of the target loss function according to the plurality of coding vectors, the category prediction values of the plurality of sample audios and the category labels of the plurality of sample audios, and updating parameters of the feature coding model based on the target loss value to obtain a trained feature coding model. The trained feature coding model obtained by the feature coding model generation method can improve the identifiability of the feature vector of the audio output and the robustness of the feature coding model.

The invention discloses an intertidal zone monitoring method based on a multifunctional radar, comprising the following steps: drawing a boundary map of land and tidal flat in radar echoes; enhancingradar echoes images; by optimizing the parameters of two thresholds T1 and T2 in the SIC inter-frame accumulation algorithm to suppress the sea clutter better, and generating a picture using the processed echo data by OpenGL; using the Canny edge detection function in OpenCV to read the current picture every three seconds and displaying a area of the current intertidal zone in real time. Accordingto the intertidal zone monitoring method based on multifunctional radar, drawing a boundary map of land and tidal flat is completed by using TinyXML technology and the OpenGL polygon separation method, the amount of calculation is greatly reduced, and a benchmark is given, which is easy to monitor; the parameters of two thresholds T1 and T2 in the SIC inter-frame accumulation algorithm are optimized to suppress the sea clutter better, the identifiability of the intertidal zone is increased; in addition, the area of the current intertidal zone can be displayed in real time, and the identification accuracy of the intertidal zone is improved.

The invention belongs to the technical field of chemical fiber industry, and particularly relates to a chemical fiber fluttering quality detection method and system based on a convolutional neural network. The system comprises an image acquisition and transmission module, a part detection positioning module, a parameter feedback module, a part recognition and judgment module, and a comprehensive judgment module. A CNN deep learning scheme is used for detecting the fiber fluttering abnormity, and a fiber fluttering abnormity detection technology of key part abnormity detection is provided. A dynamic feedback mechanism and a comprehensive evaluation mechanism are adopted; the mode of robot patrol camera snapshot and background analysis is used for replacing manual detection of abnormal conditions such as fiber fluttering and the like.

The invention relates to the technical field of photochemical detection and analysis, and discloses a fluorescent compound for detecting viscosity of oil displacement of tertiary oil recovery, and a preparation method and application of the fluorescent compound. The fluorescent probe is 2-(4-bis(4-methoxyphenyl)amino)phenyl)anthracene-9,10-dione (TPAMD). The probe compound disclosed by the invention is composed of groups with an aggregation-induced emission characteristic and anthraquinone groups with an electron-withdrawing effect, and thus has an aggregation-induced emission characteristic and an intramolecular charge transfer characteristic simultaneously. The probe is easy to prepare, is high in yield, has good viscosity responsiveness, and has relatively good linear response. The probe can be used in complex solvent environments, has good resistance to solvent perturbation, can be used to achieve detection of viscosity of chemical displacement agents in the oil production field, and the method has an analysis and detection potential for various viscosity-sensitive fields such as chemistry, food, environment and the like.

The invention discloses a decoding method of a novel layered-structure two-dimensional code. For a novel layered two-dimensional code picture, a soft decoding method is used to preferentially decode primary information and then decode secondary information. The method possesses an advantage that reduction accuracy is high and is adapt to a current rapid development demand. In the invention, through scanning the layered two-dimensional code picture, gray-scale quantification image processing is used to carry out standardization on the two-dimensional code picture which may generate distortion so as to acquire a standard two-dimensional code matrix; soft-decision demodulation is performed on the generated standard two-dimensional code matrix so as to acquire a soft information bit stream; and the soft information bit stream is input into a matrix de-interleaver decoder and a Turbo code decoder to carry out error correcting, decoding and reduction so as to acquire a source data information flow, and decoding is completed and reduction accuracy of decoding is increased.

The invention discloses a positioning method and device based on an environment image. Area features of all areas contained in the obtained environment image serve as positioning bases. Regional features represented by each region are distinctly different from features of other regions in the environment image, so that the identifiability among the regions is improved, and the phenomenon of confusion among the regions is avoided. Furthermore, each region has region characteristics obviously different from those of other regions, so that each region can be used as a positioning basis to different extents, and even if some regions in the environment image are poor in quality (such as being shielded), positioning can be realized through other regions with good quality in the environment image. Through the method and the device in the specification, the phenomenon of positioning failure can be avoided to a greater extent, and the positioning success rate, the positioning accuracy and the positioning precision can be improved.

The invention discloses an anti-counterfeit label making robot assembly line based on expansion and stretching. The anti-counterfeit label making robot assembly line based on expansion and stretchingcomprises a bottom plate, an expansion mandrel, an inner expansion and stretching mechanism as well as a stretching film roll, a primary material pulling mechanism, a particle attaching groove, a secondary material pulling mechanism, an outer expansion and stretching mechanism and a film layer slitting mechanism; the expansion core shaft comprises a feeding part, a material absorbing bending part,a conical expansion part and a forming slitting part which are connected sequentially; the feeding part faces towards the stretching film roll and passes through the primary material pulling mechanism; the material absorbing bending part stretches into the particle attaching groove; the secondary material pulling mechanism is connected between the material absorbing bending part and the conical expansion part; the inner expansion and stretching mechanism is arranged in the forming slitting part; left and right side regions of a tubular plastic stretching film of the stretching film roll are coated with adhesive layers respectively; upper and lower side regions of the tubular plastic stretching film of the stretching film roll are stretching parts without viscosity; and the anti-counterfeit label making robot assembly line based on expansion and stretching can achieve the aim of eliminating the lap joint phenomenon of the particles, can improve the identifiability of the anti-counterfeit label and has more excellent anti-counterfeit effect.

The present invention discloses a website display method. The method comprises: receiving account information input by a user in an account input interface of the website; when the account informationis input correctly, according to a mapping relationship between the account information and a webpage theme, obtaining a target webpage theme associated with the account information; and displaying the target webpage theme in the display webpage interface of the website, so that according to the displayed target webpage theme, the user can determine whether the website is phished by the phishingwebsite. The present invention also discloses a website display apparatus. According to the method and apparatus disclosed by the present invention, when the website is accessed through account information login, the user can determine that the website is phished by the phishing website if the displayed webpage theme is inconsistent with the preset webpage theme, so that the identifiability of thetarget website can be enhanced, and the risk of being phished can be effectively reduced; and the more the personalized dimensions are, the more difficultly the website can be phished by the phishingwebsite, so that the defense function to the phishing website is greatly enhanced, and the information security is improved.

The invention relates to the technical field of photochemical detection and analysis, and discloses a fluorescent compound used for oil flooding viscosity detection in tertiary oil recovery, a preparation method and application thereof. The fluorescent probe is 2-(4-(bis(4-methoxyphenyl)amino)phenyl)anthracene-9,10-dione (TPAMD). The probe compound of the present invention is composed of groups with aggregation-induced luminescent properties and anthraquinone groups with electron-withdrawing effects, and simultaneously have aggregation-induced luminescent properties and intramolecular charge transfer properties. The probe is easy to prepare, has a high yield, has a good response to viscosity, and has a good linear response. At the same time, it can cope with complex solvent environments and has good resistance to solvent interference. It can detect the viscosity of chemical displacement agents in the oil recovery field. , has the potential for analytical detection in many viscosity-sensitive fields such as chemistry, food, and the environment.

The present invention provides a print order receiving device for image data which is equipped with a slit shape taking-in / out port having inserting / drawing recording media and a plurality of read driving devices RD arranged so that taking-in / out port for recording media face one flank side of a device housing so as to read image data recorded on a plurality of kinds of recording media; and a holder part 35a which holds rolled recording paper 37 and a printer unit 35b which performs printing to the recording paper and has its vertical width smaller than the holder part are arranged side by side so that the holder part is positioned more nearly to the inside surface of the device housing than that of the printer unit and said printer unit is positioned vertically in the presence range of the holder part, said small-sized read driving device SR having small length in inner direction of the plurality of read driving devices and said printer unit are arranged closely one over the other. Taking-in / out port faces formed by all terminal edges of said taking-in / out ports are arranged in state of uptilting.

The present invention provides a print order receiving device for image data which is equipped with a slit shape taking-in / out port having inserting / drawing recording media and a plurality of read driving devices RD arranged so that taking-in / out port for recording media face one flank side of a device housing so as to read image data recorded on a plurality of kinds of recording media; and a holder part 35a which holds rolled recording paper 37 and a printer unit 35b which performs printing to the recording paper and has its vertical width smaller than the holder part are arranged side by side so that the holder part is positioned more nearly to the inside surface of the device housing than that of the printer unit and said printer unit is positioned vertically in the presence range of the holder part, said small-sized read driving device SR having small length in inner direction of the plurality of read driving devices and said printer unit are arranged closely one over the other. Taking-in / out port faces formed by all terminal edges of said taking-in / out ports are arranged in state of uptilting.

The invention relates to a twin classifier certainty maximization method for a cross-domain complex visual task; the method comprises the following steps: 1, constructing a neural network frame, and initializing the neural network frame; 2, inputting the source domain sample into a feature generator G to obtain a corresponding feature representation; and 3, under the supervision of the label information of the source domain sample, calculating an empirical risk error between a model prediction output p and a real label y on the source domain sample by using a standard cross entropy loss function. The method has the advantages that a novel classifier deterministic difference measurement CDD is designed, wherein the difference of the classifiers is measured by means of category correlation between target prediction of the twin classifiers, and implicit constraints are applied to the identifiability of target features at the same time.

The invention relates to a small sample behavior recognition method based on multi-dimensional prototype reconstruction and reinforcement learning. The average prototype of the class, and the re-weighted similarity attention is used to calculate the similarity between the query set sample and the support set sample and the class average prototype, and re-weight the support set sample and the query set sample according to their corresponding similarity, and get two Prototype, the two prototypes are weighted and summed to obtain a cross-enhanced prototype, and a double triplet optimization classification feature space is constructed to enhance the discriminative ability of the cross-enhanced prototype for different categories, and the optimized cross-enhanced prototype is used to identify all categories. The video in the query set sample is classified, which greatly improves the classification accuracy.

The invention discloses an image identification method of stirring friction welding arc stripe interval. The image identification method is characterized by mainly comprising the steps of: obtaining image, carrying out binarization processing on the image, carrying out post processing on the binary image and detecting and calculating the arc stripe interval, firstly, obtaining a microscopic picture of a stirring friction welding arc stripe tool; and then, performing gray scale, wave filtering and binarization processing on the microscopic picture of the arc stripe tool; removing the isolated points, the interior stripe detects and the incomplete edge stripes of the binary image; and finally, determining the arc stripe interval calculation range, calculating the arc stripe interval and processing the data. The image identification method of stirring friction welding arc stripe distance disclosed by the invention greatly enriches the obtaining method of arc stripe parameters and provides certain foundation for research of the arc stripe parameters in the stirring friction welding.

The invention discloses a medical image information marking method and device and a medical device. The medical image information marking method includes the steps of distributing a unique examination serial number to each patient on whom the medical image examination is conducted, and marking medical images of the patients with the examination serial numbers. According to the technical scheme, the identifiability of image data of the patients can be improved.

The invention relates to a design method of a linear discriminative sparse representation classifier based on kernel space, the steps are: read training samples, perform nonlinear transformation on the training samples, transform them into high-dimensional kernel space, and perform each A class of training samples is used for learning, and the contribution (ie weight) made by each individual in this class of training samples to construct the subspace of this class of training samples is found. The product of this class of training samples and the weight matrix forms a dictionary, and all categories of The dictionaries are arranged in turn to form a large dictionary matrix; the linear discriminant sparse code of the test sample in the kernel space is obtained through the dictionary matrix, and the test sample is fitted with each type of dictionary and the linear discriminant sparse code corresponding to the dictionary, and the fitted Error; the class with the smallest fitting error is the class of the test sample. The invention can ensure that the sparse codes of samples of the same type are concentrated, and the sparse codes of samples of different types are dispersed, thereby effectively increasing the discrimination of samples and improving the performance of classifiers.

A car roof trim cover integrated assembly comprises a front top lamp trim part, a top cover left trim strip, a top cover right trim strip, a front top daytime running light, a top cover rear identification part and a top rear trim part. The front ceiling lamp decorating part is arranged at the front part of the roof of the commercial travel vehicle; the top cover left decoration strip is positioned on the left side of the roof; the front end of the top cover left decoration strip is connected with the left side surface of the front ceiling lamp decoration part; the front end of the top cover right decoration strip is connected with the right side face of the front top lamp decoration part, the front middle portion of the front top lamp decoration part is provided with a backwards-inclined air guide face from bottom to top, lamp grooves are formed in the positions, close to the lower portion, of the two sides of the air guide face, and front top daytime running lamps are arranged in the lamp grooves; top cover rear identification pieces are embedded in the outer sides of the rear ends of the top cover left decoration strip and the top cover right decoration strip. The top rear decorative part is arranged on the rear portion of the roof, and the two ends of the top rear decorative part are connected with the inner sides of the rear ends of the top cover left decorative strip and the top cover right decorative strip respectively. The space in the vehicle can be expanded, the influence of wind resistance is reduced, the identifiability of the vehicle is improved, accidents are prevented in advance, and driving safety is guaranteed.

A structure of vehicular illuminator can ensure enough brightness of bulb light transmitting to different directions and can execute reliable light distribution in the structure of vehicular illuminator which executes light distribution in the direction different from the reflection direction of main reflector. The structure of vehicular illuminator is provided with the following components: a notch (81) which is formed on a main reflector (67) and guides the bulb light to the back side of main reflector (67), and an auxiliary reflector (82) which extends to the back part of main reflector (67) along the irradiation direction of bulb light guided by the notch (81). The bulb light which passes through the notch (81) and is guided to the back part of main reflector (67) is executed with light distribution in a direction different from the reflection direction of main reflector (67) by the auxiliary reflector (82).

The invention discloses a device for automatically weighing and separating sheep flock and a using method of the device, and relates to the technical field of novel automatic sheep flock separating devices. The device comprises a grouping carrying outer frame, the top end of the grouping carrying outer frame is fixedly connected with a PLC control cabinet, and one end of the grouping carrying outer frame is fixedly connected with an automatic unfolding entrance door. According to the invention, through design of a synchronous logarithmic automatic marking identification structure, the device is convenient to automatically label bodies of classified sheep again and shoot and record cloud data, so that the recognizable degree after classification is further improved, and the convenient support for number modularization is improved; and through the design of the adaptive guide adjusting plate structure, the device can conveniently complete adaptive gradient and angle driving adjustment, so that the device adapts to leg heights of different species, better stress support is formed, and the device can conveniently and better guide different livestock to use the device.