118results about How to "Improve identification accuracy" patented technology

A prescription verification system includes a database that contains a plurality of spectral signatures corresponding to identified pharmaceuticals. A multimodal multiplex sampling (MMS) spectrometer obtains a spectra of a pharmaceutical to be identified and verified. The pharmaceutical can be inside or out of a vial. The prescription verification system includes algorithms for matching spectra of pharmaceuticals to be verified obtaining using the MMS spectrometer to spectral signatures contained in the database corresponding to identified pharmaceuticals. The prescription verification system further includes algorithms for identifying such pharmaceuticals to be verified.

The invention provides a spectrum library generating method, comprising the following steps: selecting analyzed spectrograms of an experimental tandem mass spectrometry, in which information of female ionic peptide sequences, electronic charges, modified types and sites are included; removing redundant spectrograms from the analyzed spectrograms of the experimental tandem mass spectrometry to obtain representing spectrograms; dividing the female ionic peptide sequences corresponding to the representing spectrograms according to a theoretical fragmentation mode to obtain theoretical spectrograms corresponding to the representing spectrograms; combining the theoretical spectrograms and the representing spectrograms to obtain optimized spectrograms; and labeling spectral peaks for the optimized spectrograms, and utilizing the optimized spectrograms with labeled spectral peaks to generate a spectrogram database. The invention further provides a spectrogram identifying method of the tandemmass spectrometry. In the invention, in the process of matching candidate spectrograms and the spectrograms of the tandem mass spectrometry to be analyzed, mass-to-charge ratio offset of the spectralpeak possibly introduced by a potential modification is concerned so that ionic spectral peak containing modified fragments is matched, thereby achieving better modified spectrogram identifying effect.

The invention discloses a noise motion blurred picture recovery method based on radial neural net, which comprises the following steps: (1) making the noise motion blurred picture y(m, n) generate the corresponding smoothing picture s(m, n) after low-pass filtering with the two-dimensional median filter; (2) calculating the inaccuracy picture e(m, n)=y(m, n)-s(m, n) between the noise motion blurred picture and the smoothing picture; (3) detecting the edge with Canny operator in order to estimate the gradient f' (m, n) of y(m, n); (4) calculating regular parameter lambda (m, n) of every picture element according to the size of f' (m, n) and generating interpolation picture f lambda (m, n) by radial neural net RBFN according to the inaccuracy picture e(m, n); (5) acquiring the silent motion blurred picture f(m, n) by superimposing the interpolation picture f lambda (m, n) to the smoothing picture s(m, n) after low-pass filtering; (6) acquiring the width HL and the height VL of two-dimensional blurred picture by identifying the motion blurred direction and the motion blurred length of the motion blurred picture automatically and acquiring the recovery picture with the picture recovery method. The invention provides the automatic identification, the low complicated calculation and the good recovery effect.

The invention provides an abnormal behavior early warning method and system for hovering event space-time big data analysis. The method comprises the steps: according to the hovering behavior alarm information sent from intelligent monitoring cameras deployed by a monitoring point, establishing an alarm big data event base so as to carry out correlation analysis of the real time and historical big data for the alarm events for a single point and the space-time big data correlation analysis among the alarm events for a plurality of monitoring points and to deeply excavate the risk grade of the robbery criminal behavior concealed in the alarm events. For conveniently identify whether different times of hovering footprinting behaviors are made by the same people or the same group of people, the invention specifically provides a hovering pedestrian re-inspection algorithm. On the premise of not adding extra hardware cost for the monitoring system, the abnormal behavior early warning method and system for hovering event space-time big data analysis overcome the defect that single time of alarm information of the single intelligent camera is not enough to be the early warning basis, thus effectively improving the efficiency of the intelligent monitoring system.

The invention belongs to the technical field of computer vision, and particularly relates to a labor protection article wearing condition detection and identity recognition method based on deep learning. The method comprises the following steps of collecting data to make a special data set for worker labor protection article wearing condition detection and identity recognition; preprocessing, labeling and dividing the data set; designing and training each depth model based on multiple target detection algorithm principles; carrying out model integration on a plurality of depth models by adopting a Stacking Enable method; and training a face recognition model and a pedestrian re-recognition model based on deep learning by using the public data set, constructing a face recognition and pedestrian re-recognition comparison database of a work site, performing adaptive optimization on the models in combination with the image database, and performing detection and monitoring. The safety of workers in a construction site is improved, safety accidents are reduced, good social value and commercial value are achieved, and the application prospect is wide.

The invention belongs to the technical field of computer auxiliary diagnosis, and specifically relates to an identification method of primary central nervous system lymphoma and glioblastoma based on a sparse representation system. The method includes: segmenting T1 enhanced and T2 weighted MRI image tumor regions by employing an image segmentation method based on a convolutional neural network; then designing a dictionary learning and sparse representation method, and extracting texture characteristics of the tumor regions; selecting some characteristics with high stability and high resolution for tumor identification by employing an iterative sparse representation characteristic selection method in order to reduce the characteristic redundancy and improve the tumor identification efficiency; and finally establishing a combined sparse representation classification model containing two modals of T1 enhanced or T2 weighted based on the thought of eigenstate fusion in order to improve the tumor identification precision. According to the method, high tumor identification precision can be obtained, manual operation for extraction of identification parameters is avoided, the robustness is high, and the method can be applied to clinic identification of primary central nervous system lymphoma and glioblastoma.

The invention belongs to the technical field of social network user identification, and discloses a cross-social network user identity identification method based on a community structure, which comprises the following steps: step a, respectively mapping a source network and a target network to a low-dimensional vector space by adopting a network embedding mode; and step b, based on the vector space, training a BP neural network in a supervised manner to obtain a BP neural network model for realizing mapping from a source network to a target network, and identifying the user identity of the source network in the target network through the BP neural network model. In the process of learning the feature vector representation of the social network nodes, the proximity features and the community structure features of the nodes are fused, the structure features of the social network are reserved to the maximum extent, and the user identity recognition accuracy is improved.

A highly accurate identification system is built by using an identification module, which has predetermined values of FAR and FRR. A plurality of (m) biometrics are registered and managed. At the time of identification, a collation process is executed by using the m biometrics, and also a collation process is executed a plurality of (n) times with each biometrics. An identification result of the system is judged and outputted by combining the results of the m-biometrics collation and n-time collation trials. As conditions of judgment, when the collation is judged to be successful even once in the n-time collation trial, the collation is judged to be successful with respect to that biometrics. In the in-biometrics collation trial, when the collation is judged to be successful with respect to all the m biometrics, a judgment of identification success is given in the system as a whole.

The invention discloses a hyperspectrum-based transgenic maize kernel recognition device and method. The method comprises the following steps that crylAb/cry2Aj-G10evo transgenic maize and nonparent maize hyperspectral images are obtained through a hyperspectrum instrument; spectral information of an area of interest is obtained through an image processing method, and spectroscopic data preprocessing is performed; all spectra and extracted characteristic spectra are used for building a PLS and SVM discriminant analysis model. The recognition rate of the built transgenic maize kernel SG-PLS model is 90% or above. Compared with a traditional molecular detection method, rapid, visual and lossless identification of transgenic maize and non-transgenic maize kernels can be achieved, operation is easy, no large amount of experimental treatment is needed, the cost is low, and the good application prospect and considerable market value are achieved.

The invention relates to a method for identifying Chinese herbal medicine by using THz-TDS (terahertz-total dissolved solids) in combination with a fuzzy rule expert system. The method mainly comprises the following steps: detecting the Chinese herbal medicine sample by using THz-TDS to obtain a terahertz time-domain spectroscopy, pretreating the terahertz time-domain spectroscopy, and dividing the terahertz time-domain spectroscopy to obtain a training set sample and a verification set sample, wherein the fuzzy rule expert system is used by the training set sample and the verification set sample to establish a qualitative analysis model so as to identify the Chinese herbal medicine sample. The method provided by the invention is simple and pollution-free, and easily realizes the automation; the Chinese herbal medicine can be fast and accurately identified in lossless, and the method is suitable for the quality control in the production process of the Chinese herbal medicine.

The invention provides a method and a device for attendance of persons descending a well on the basis of face recognition. The method comprises the steps of automatically segmenting image areas of all the persons by analyzing pictures shot for the persons descending the well, then carrying out face detection on all the image areas, and finally determining the identities of the persons by using a face recognition technology so as to obtain the well descending condition of the persons descending the well. The method and the device provided by the invention have the advantages that the traditional name-calling and answering mode is avoided, the valuable time is saved, and simultaneously the evidence is left for performance evaluation and analysis.

The invention relates to the technical field of campus teaching management, in particular to a teaching management system and method based on face recognition, the method comprises a server and a camera device, the server comprises a database, and the server is used for recording face information of students in a whole school and storing the face information in the database; the server is used forcalling the face information of the students in the whole school in the database and carrying out face recognition on all the students in the classroom through the camera device; the server is used for marking the students passing the face recognition and recording the marked students as successful sign-in; and the server is used for counting the class arrival rate, the class listening rate and the activeness of the students in the classroom through face recognition. The system comprises the following steps: a face information input step, a face recognition step, an automatic sign-in step, aclass condition analysis step and a class condition evaluation step. The system and method overcome the defects that a traditional attendance checking mode is long in attendance checking time, replaces answering and the like.

The invention discloses a method for identifying different transgenic rice. The method comprises the following steps of (1) emitting near infrared spectrums to rice seed samples, and collecting diffuse reflection spectral information of all the rice seed samples; (2) respectively preprocessing diffuse reflection spectral information of all the rice seed samples, extracting spectral characteristic information of a characteristic spectrum section subjected to the preprocessing through a principal component analysis method, and selecting principal components to obtain scores of the principal components; (3) establishing a model with the scores of the principal components corresponding to the spectral information of all the rice seed samples serving as input and transgenic type set values corresponding to the rice seed samples as output; and (4) obtaining scores of the principal components of spectrums of rice seeds to be measured, and putting the scores of the principal components into the model obtained in the step (3) to obtain transgenic types of the rice seeds. Trans-protein-gene rice and trans-controlling-gene rice are utilized in the method. The method is simple in operation, high in identification accuracy degree and capable of achieving fast and nondestructive identification of different transgenic rice.

The invention discloses a banknote processing device, including a control part, a storage part, an operation part, a transportation part, a reading part, a judging part and an identifying part which are connected through a bus. The reading part reads banknote characteristic quantities and depreciation state; the storage part stores three sets of banknote characteristic quantity reference information of different threshold ranges, each set includes a plurality of characteristic quantity reference parameters, and each reference parameter has a corresponding threshold maximum value and minimum value; the judging part calls the banknote depreciation state, and comprehensively calculates the banknote depreciation degree; the identifying part calls the characteristic quantity reference information of the corresponding level according to the banknote depreciation degree, and compares each characteristic quantity of a banknote with the corresponding threshold range, and if all the characteristic quantities are located in the corresponding threshold range, the banknote is identified as a genuine banknote; and the control part controls working states of all components. The technical scheme improves precision of identifying authenticity of banknotes, and avoids the problem that possibility of mistakenly identifying forged banknotes, incomplete banknotes and stained banknotes as genuine banknotes becomes high.

The embodiment of the invention provides a multi-spectral recognition device for banknotes, which is used for acquiring images of a banknote in each optical spectrum section, comparing the images with preset conditions and improving the recognition accuracy. The technical scheme is that the multi-spectral recognition device for banknotes provided by the embodiment of the invention comprises a wide field imaging module, a wavelength adjusting module, a photoelectric detection module and an identification module; the wide field imaging module is used for acquiring a reflection image of the banknote to be detected; the wavelength adjusting module is used for extracting n reflection sub-images of the reflection image under n optical spectrum sections, wherein the n is a natural number which is greater than 2; the photoelectric detection module is used for carrying out photoelectric signal conversion on the n reflection sub-images, and acquiring n converted images corresponding to the n reflection sub-images; and the identification module is used for recognizing the authenticity of the banknote to be detected through the n converted images.

The invention relates to an optical detection method. The method is used for detecting the truth of color-changing ink of paper money or a holographic image. The optical detection method comprises the following steps: an image to be detected is arranged in an effective range capable of being irradiated by a light source; the light source and a camera begin to work; the light source irradiates the image to be detected; the ray of a part of the image to be detected, irradiated by a light beam inclining on the image to be detected, enters an oblique camera after being reflected by at least one reflector; the ray vertical to the image to be detected enters a vertical camera to shoot, and enters to be parallel to the ray of the oblique camera and the vertical camera, and the shot picture is analyzed. An illuminating system comprises a light source, and at least one reflector arranged on a light path from the light source to the oblique camera or the light path on which the vertical camera is; an imaging system comprises at least one vertical camera and the oblique camera, and is arranged at one side of the illuminating system. The optical detection method is simple and flexible in using method, high in identification accuracy comparison and low in cost.

The invention relates to a time-resolved oil fluorescence identification method based on main component analysis. The time-resolved oil fluorescence identification method comprises the following steps of: carrying out feature extraction on an oil fluorescence spectrum through a main component analysis process; establishing vector space by utilizing extracted main component, carrying out correlation analysis among samples to obtain an oil fluorescence time sequence feature sequence, and calculating the maximum value and dispersion parameters of the feature sequence; and respectively carrying out hierarchical clustering to obtain types of oil according to the maximum value and dispersion parameters of the feature sequence. According to the time-resolved oil fluorescence identification method based on main component analysis disclosed by the invention, feature change sequence consisting of oil fluorescence main component characteristics at different sampling moments is further analyzed, and the maximum value and dispersion parameters used for time-resolved oil fluorescence feature parameterization are introduced, so that number of petroleum product types capable of being identified is increased, and petroleum product type identification ability is improved.

The invention provides a voiceprint recognition and face recognition combined seal control instrument application method and system, and relates to the field of seal safety use. The method comprises the following steps: acquiring an audio template and a face image template of a person authorized by a seal control instrument; when a seal control instrument user uses the seal control instrument, collecting audio of the user for audio authentication; after authentication, collecting a face image of a printing control instrument user, representing the image and a face image template by using sparse coding, and calculating the similarity between the face image of the user and the face image template; using a saliency detection method for carrying out saliency detection on a face image of a user and a face image template, carrying out sparse coding representation on a saliency region, and calculating the similarity of the saliency region of the face image and the face image template. A strategy of combining voiceprint recognition and face recognition is utilized, so that the identity recognition precision of a user of the printing control instrument is improved, and the safety coefficient of the printing control instrument is increased; the used method is classic and simple, does not need to consume more computing resources, and can reflect better real-time performance.

The invention relates to a method for identifying Chinese agaricus sinodeliciosus. The method includes the steps of (1) external form identifying; (2) microstructure identifying; (3) chemical reagent detection; (4) molecular sequence identifying. According to the method, the A. sinodeliciosus is identified with the method of combining identifying of the external form characteristic, changes in the color of an anatomical structure, a chemical reagent and the molecular sequence, and the method has the advantages of being high in identification accuracy and accurate in positioning; the method is representative in identification of Chinese agaricus sp. classes, and has good application prospects.

The invention provides a natural shoreline automatic acquisition method based on laser point cloud data and an average large tide surface. The natural shoreline automatic acquisition method comprises the following steps: acquiring laser point cloud data from land to an intertidal zone when an ocean ebbs; acquiring water level data, and calculating a plurality of years of annual and overall average large-tide high-tide surfaces and respective corresponding elevation values of the annual and overall average large-tide high-tide surfaces; if the difference result between the laser point cloud data and the annual average high-tide elevation value is within the laser point cloud error range, taking the laser point cloud data as natural shoreline association point data; and processing the natural shoreline association point data to obtain natural shoreline data, predicting the future trend, and analyzing the correlation between the periodic change of the natural shoreline data elevation and the coast erosion. The method is less intervened by investigators, is high in automation degree, and improves the working efficiency and the natural shoreline determination precision; as the water level data is real-time water level observation data or water level data forecast through tide, the water level data is high in current potential, the calculated average high tide surface is high in precision, and the natural shoreline extraction precision is high.

The present invention is to improve the detection and the identification accuracy of an object in image analysis in a camera control system that controls a camera used to take an image to be analyzed by artificial intelligence. The image analysis system that performs machine learning by using a plurality of teacher data associating a label that indicates what the object is with image data to which the label is attached, includes an imaging condition analysis module 211 that analyzes the imaging condition of teacher data; and a camera control module 212 that controls a camera to take an image under the analyzed imaging condition, to analyze the image taken under the imaging condition.