174results about How to "Guaranteed similarity" patented technology

The invention proposes a method and apparatus for significantly improving the quality of voice morphing and guaranteeing the similarity of converted voice. The invention sets several standard speakers in a TTS database, and selects the voices of different standard speakers for speech synthesis according to different roles, wherein the voice of the selected standard speaker is similar to the original role to a certain extent. Then the invention further performs voice morphing on the standard voice similar to the original voice to a certain extent, in order to accurately mimic the voice of the original speaker, so as to make the converted voice closer to the original voice features while guaranteeing the similarity.

The invention discloses an image retrieval method based on variable-length depth hash learning and mainly relates to the field of image retrieval and depth learning. According to the method, learning of hash codes is modeled into the process of similarity learning. Specifically, the method utilizes a training image for generating a batch of ternary image sets. Each ternary image set comprises two images with the same label and one image with the different label. The purpose of model training is to space image pairs matched to the maximum and the unmatched image pairs in hamming space. The depth convolution nerve network is introduced into the learning part of the method, the image characteristics and hash functions are optimized in a combined mode, and the end-to-end training process is achieved; hash codes output by the convolution network have different weights. For the different retrieval task, a user can regular the length of the hash codes by disconnecting unimportant bits; meanwhile, according to the method, the discrimination of the hash codes can be kept effectively under the circumstance that the hash codes are short.

The invention relates to a human body gastrointestinal tract bionic system and a simulation experiment method based on the system. The system comprises a system controller, an intestinal flora planting system, a simulation digestion system, a simulation absorption system, a stirring system, a rate control system, a gas path system, a temperature adjusting system, a pH adjusting system, a digestion liquid reagent bottle, an absorbing liquid reagent bottle and a collector. The method comprises the following steps: setting parameters of the control system, preparing planting colloidal particles, planting the intestinal flora, simulating digestion and the like. The system simulates gastrointestinal tract digestion, absorption and intestinal microecology of human bodies under in-vitro conditions, is real in simulation and good in reproducibility, and us applicable to study on the fields of nutriology, toxicology, physiology, microbiology and the like as intelligent control is adopted.

The invention discloses a method for reconstructing human facial image super-resolution based on the similarity of facial characteristic organs. The method comprises the following steps of: 1, establishing a high-resolution front human facial image library and a high-resolution characteristic organ image library by utilizing a gray scale projection method according to a preset ideal high-resolution human facial image; 2, extracting a low-resolution characteristic organ image from a low-resolution target human facial image; 3, performing bicubic interpolation on the low-resolution target humanfacial image and the low-resolution characteristic organ image to acquire a training image set of the low-resolution image; 4, constructing characteristic space corresponding to the training image set by the training image set to reconstruct projection vectors of a corresponding high-resolution integral human facial image and a corresponding high-resolution organ image; and 5, fusing the high-resolution integral human facial image and the high-resolution characteristic organ image into a high-resolution target human facial image. The method has the characteristics of less preprocessing time, high retrieval accuracy of training images, high trueness of the acquired human facial images and the like.

The invention relates to a method and device for ceramic denture 3D printing suitable for light curing molding and belongs to the technical field of tooth making. According to the method, through scanned tooth data and data processing, data of each layer of image after a data model to be printed is sliced are obtained and transmitted to a control system, a liquid supplementing action is controlled through the control system so that the liquid level of a base material can be always kept at the preset height, the base material is configured by photosensitive resin and ceramic powder according to the ratio, and the control system controls completion of stirring of supplemented liquid after liquid supplementing; and a bottom projection manner is adopted for printing, slice images of each layer of image are projected successively, light source exposure is conducted, and curing is completed layer by layer. The method and device have the beneficial effects that the making process is simple, the making efficiency is high, the consumption of materials is low, the making precision is high, and the match is good.

The invention relates to an automatic generation method of an interactive three dimensional city panoramic map. Based on city morphology and gestalt theory, the method combines roads widening, viewpoint raising, building displacement and building height reducing with optimization method through establishing energy function deviating argument and shielding argument for estimating city space cognition and uses linearization constrained condition energy function to solve minimizing energy function in real time, therefore automatic generation of the interactive three dimensional city panoramic map is achieved. The making method of the three-dimensional digital city panoramic map is characterized by fast visualization speed and little deformation of city scenes, having broad application prospect in city navigation, space position service and the like.

The invention provides a large-scale mixed heterogeneous storage system-oriented node fault prediction system and method. A time sequence-based association rule mining algorithm is adopted to construct a node fault prediction system architecture, and a main process of node fault prediction includes: collecting state data and log information of each storage node; carrying out data preprocessing, and generating sequence modes on the basis of a sliding window; using the sequence modes and fault sequences, which are extracted in a fault identification process, together as input of an association rule algorithm, and outputting output results as typical fault sequences; carrying out matching on the typical fault sequences and sequence modes generated in real time; and if a matching result meetsan established rule, issuing early warning to notify a system administrator, and giving feedback to a prediction result by the administrator according to a subjective interest degree. According to thesystem and method, real-time online fault prediction is carried out for nodes of a large-scale mixed heterogeneous storage system, and accuracy and recall which are better than those of existing fault prediction algorithms and better scalability can be obtained.

The invention relates to an apparatus and method for measuring elastic turbulence characteristics of a viscoelastic polymer in a porous medium, wherein a side connector of a fluorescent microscope of the apparatus is connected to a CCD (couple charged device) camera, a saline container, a polymer solution container and a synthetic oil container that are connected with an oil / water container are parallelly connected between an inlet of a micro metering pump and an inlet of a two-dimensional mesh microscopic simulation model, the two-dimensional mesh microscopic simulation model is provided with an inlet pressure sensor and an outlet pressure sensor, and a light guide arm of a micro field particle image speed-measuring system is positioned right above the two-dimensional mesh microscopic simulation model; all the micro field particle image speed-measuring system, the fluorescent microscope, the inlet pressure sensor and the outlet pressure sensor are connected to a data graphic data station. The apparatus and method are capable of effectively recognizing elastic turbulence of viscoelastic polymer solution in porous medium seepage, and providing qualitative and quantitative description for elastic turbulence characteristics.

The invention relates to a face super-resolution reconstruction method based on an identity priori generative adversarial network. The face super-resolution reconstruction method comprises the following steps: firstly, reading an original face picture data set; then, utilizing a face image-identity label pair to train a face feature extraction network; thirdly, reading the high-resolution face image for bicubic interpolation down-sampling to obtain a high-resolution face image-low-resolution face image pair for model training; fourthly, inputting the low-resolution face image into a generatornetwork to generate a super-resolution face image; respectively inputting the high-resolution face image and the super-resolution face image into a trained face feature extraction network, and extracting identity prior features of the high-resolution face image and the super-resolution face image; and inputting the high-resolution face image, the super-resolution image and the corresponding identity prior features into a discriminator network, calculating a supervised adversarial loss function by using the output of the discriminator network, and training a generative adversarial network by using error back propagation.

The invention relates to the technical field of touch recognition, and provides a mobile terminal and a method and system for getting access to application programs in a ready mode. The method comprises the following steps of obtaining coordinate data, recognizing gestures and getting access to the application programs. In recognition, similarity recognition and distance recognition are combined to be matched with a user-customized gesture template, similarity of the gesture template is guaranteed according to a gesture overall contour and a gesture specific position, and therefore recognition accuracy of a complex gesture can be improved. Moreover, gestures which are customized by the user according to personal preferences or habits and used for unlocking and getting access to the application programs are supported, the gestures are not limited by an existing sudoku and can be any characters or words or figures, and usage experience of the user is improved.

The invention discloses a test device for simulating a steam generator secondary side waste heat exhaust system and a method. The test device comprises a heat source simulation body, a cold source simulation body, a final heat sink heat exchanger, a final heat sink water tank, a first pipeline and a second pipeline, wherein the heat source simulation body and the cold source simulation body are connected through the first pipeline to form a first loop, the final heat sink heat exchanger is arranged in the final heat sink water tank, the cold source simulation body and the final heat sink heat exchanger are connected through the second pipeline to form a second loop, the first loop has a first heat exchange medium, the second loop has a second heat exchange medium, the heat source simulation body heats the first heat exchange medium, and the first heat exchange medium realizes heat exchange with the second heat exchange medium in the cold source simulation body. According to the test device, the first and second loops of a reactor are taken as prototypes, an H2TS method is utilized for simulation, integral phenomena and local phenomena of the steam generator secondary side passive waste heat exhaust system can be better simulated, a water-water heat exchange process generated between the first and second loops of the reactor can be accurately simulated, and requirements are satisfied.

The invention relates to a robot demonstrating track generating method based on time-space characteristics. The robot demonstrating track generating method includes steps of 1) recording demonstrating tracks of various degrees of freedom at the tail end of a robot at the same sampling time interval by operating the robot and demonstrating the like tracks multiple times; 2) extracting the common time-space characteristics from the various same types to generate a track template; 3) setting up a target equation meeting the time-space characteristics, and solving the motion track of the robot by a covariant gradient descent method; 4) utilizing the motion track as the reference track to input into an executing mechanism of the robot, tracing the tracks by a controller of the robot so as to enable the generated tracks to be consistent with the reference track. Compared with the prior art, the robot demonstrating track generating method is added with the time-space characteristics obtained by learning of the demonstrating tracks on the basis of complying with original constraints; by adding the constraints, search space of the tracks can be limited and solution speed is accelerated.

The invention relates to a multimedia binary coding method based on supervised multi-perspective discretization. The method comprises the following steps that: S1: assuming a training set formed by npieces of images, obtaining a Hash function which contains one series of sub-functions through learning, mapping two different modals of features of a sample to an optimized feature space, obtaining one series of Hash values corresponding to Hash sub-functions, and then, converting the Hash values into a binaryzation Hash code through binary quantization; S2: obtaining a Hash function based on supervised training: defining a linear multi-classification model, optimizing the model function, and adopting a minimum squaring loss as a target function; S3: obtaining a Hash function based on a minimum quantization loss: assuming the feature of one modal, and optimizing to the minimum quantization loss through the Hash function; S4: obtaining a Hash function based on a multi-perspective anchor chart: constructing the anchor chart, and adopting an anchor chart regularization Hash function; and S5: optimizing an algorithm. By use of the method, the similarity of data in an original space can bekept, and retrieval accuracy can be improved.

A method is used for measuring flow injected into a section. A tracer with polymerization characteristic is released into the oil pipe of a flooding well. A part of the tracer goes through a water faucet, flows into the ring space of an oil sleeve and enters an injection layer when passing by an eccentric distributor. When the tracer is released into the oil sleeve, an instrument with two detectors should be placed at a position through which the tracer flows. When the tracer flows through two detectors in the ring space of the oil sleeve, a reciprocal change signal will be measured by the upper and the lower detectors. The other part of the tracer continues to flow down in the oil tube and the change signal is also measured by the upper and the lower detectors. Two wave crests appear on a well log. Owing to the good polymerization characteristic of the tracer, minor fluid state change combined, the small area of the ring space of the oil tube, short going-through time and the narrow width of the crest, the inner crest value of the oil tube and the crest value of the ring space of the oil sleeve can be determined. As the inner crest value of the oil tube is removed, the time when the tracer goes through the upper and the lower detectors can be worked out. Because the section of flow speed is known, the monolayer injection flow can be worked out according to an arithmetic mode.

The invention provides a cross-modal video retrieval method and system based on a multi-head self-attention mechanism and a storage medium, and the cross-modal video retrieval method comprises a videocoding step, a text coding step and a joint embedding step. Semantic information in the training multi-modal data is fully utilized for training, a multi-head self-attention mechanism is introduced,fine interaction in videos and texts is captured, key information of the multi-modal data is selectively concerned to enhance the characterization capability of the model, data semantics are better mined, and the invention has the advantages of being high in practicability and easy to popularize. Consistency of the distances of the data in the original space and the shared subspace is ensured. Theinvention has the beneficial effects that experiments prove that the similarity of the data in the original space can be effectively maintained, and the retrieval accuracy can be improved.

The invention discloses a book recommending method based on multi-view hash, comprising the following steps of 1) screening behavior data of a user on two views from a log collection system, wherein the behavior data comprise book clicking data and search data; 2) establishing user feature vectors of the user for clicking and searching the views; 3) obtaining a user hash code, a hash function and weights of the two views through a multi-view hash algorithm by utilization of the behavior data of the two views; 4) searching a similar user for a target user by utilization of the obtained user hash code; 5) obtaining a book set clicked by the similar user, taking the book set as a recommended candidate list, calculating the book preference degree of the target user, and returning front N books with maximal preference degree of the target user. According to the book recommending method based on the multi-view hash, on one hand, the behavior data of the user on two views can be integrated into the hash code, so that the book recommendation accuracy is improved; on the other hand, the Hamming distance calculation speed of the hash code is very fast, so that the book recommendation efficiency can be improved.

The invention relates to a content erotic image zooming method based on conformal energy. The method comprises the following steps: S1. receiving an image to be zoomed, and dividing the image into grids; S2. extracting a control dot set representing local and integral characters according to the divided grids; S3. establishing an optimized energy equation according to similar constraint of the control dot set before and after being changed; and S4. using the preset image dimension as the border condition to obtain final positions of control dots by the optimized energy equation and obtain thefinal image by difference value. When the content erotic image zooming method is used for zooming the image, the image can be adjusted to any dimension and aspect ratio, and meanwhile, the distortionof important objects in the image is minimized.

The invention provides a fractured horizontal well pattern model and a method for manufacturing the same. The method comprises the following steps: selecting a rock core from natural outcrop rock cores; cutting the selected rock core into a preset designed shape, and arranging a vertical well and a horizontal well on the rock core; forming a crack in the rock core; filling an artificial rock core, which has a penetration rate which is determined according to the dimensionless flow guiding capability of the crack, in the crack; filling an artificial rock core, which has a penetration rate which is determined according to the pressure drop of the horizontal well, in the horizontal well.

The invention relates to a transmitting module of a correlated imaging laser radar. The transmitting module is characterized in that the transmitting module comprises a speckle light source, an imaging system, a data memory and a synchronous controller, wherein the speckle light source and the imaging system are fixed on a same base. The transmitting module disclosed by the invention has the characteristics of high pulse transmitting speed, high utilization rate of light energy, capability of effectively reducing ghosting strength of a reference light field, and the like.

The invention discloses a medical image retrieval method based on feature extraction and similarity matching. In the process of a radiotherapy plan, image registration is a very important key step. The purpose of the image registration in the radiotherapy plan is to find out a template image the most appropriate for a target image and to obtain a best registration result by matching operation for clinical target delineation, organ dose simulation or treatment. Therefore, it is quite important how to search out the most appropriate template image. In the method, 10 morphological features related to the form is combined with medical history information, the features are endowed with different discrimination weights, and finally 10 groups of the most similar images and related information are selected and provided for the doctor to choose, and the most appropriate template image is selected by comprehensive consideration. The operation of the method can guarantee the similarity between the template image and a target image so that it is possible to improve the applicability and the accuracy of the image registration operation in the radiotherapy plan.

The invention discloses a wave simulation generating device under a high gravity field. The adoption of a wave absorbing unit can enhance the wave absorbing effect, reduce the influence of a reflectedwave on a waveform, and ensure the similarity of a wave simulation experiment to real environment simulation. The wave absorbing unit can adjust the position of a wave absorbing plate and the size ofa grating opening according to the waveform generated by a wave generating unit so as to ensure an optimal wave absorbing effect. The wave generating unit realizes force transfer and the decoupling of motion in a simple structure. Compared with a transmission structure with a long transmission path and a complicated structure, the wave simulation generating device is less affected by mechanism jamming caused by deformation of a centrifugal field and ensures the rigidity and stability of a transmission mechanism more easily. Different from a conventional way of moving a push plate as a whole,only the upper part of the push plate is pushed, and so the load on the push plate is small and required power is small. The wave simulation generating device under a high gravity field is suitable for simulation of waves with high water depth and large amplitude.

The invention discloses a single image hair reconstructing method based on matching of a direction field and helixes. The method involves four stages. In the first stage, through user interaction division of an input single image, a head model is reconstructed by utilizing an obtained face area. In the second stage, by utilizing a hair area, reliable hair direction field information is obtained and a two-dimensional hairline segment hair model is generated. In the third stage, helix model matching is performed on two-dimensional hairlines according to the result of the second phase and the depth value of the hairline is estimated through combining with the reconstructed head model. In the fourth stage, by utilizing the two-dimensional hairline segment hair model and the corresponding depth value, an initial visible area three-dimensional hair model is obtained, an invisible area hair model is obtained by utilizing a hair database and a complete dense hair model is obtained through regrowth via the three-dimensional direction field. The invention can reconstruct the hair model which is highly similar to an original image and has characteristics of simple interaction, convenient user operation and the like.

The invention provides an expansive soil direct shear test device and a shear strength measurement method. The device comprises a strain type direct shear instrument and a loading plate, wherein the loading plate is provided with a limiting structure and a pressure sensor; a larger overlying load is applied before a sample is soaked in water, the limiting structure controls the bottom surface of the loading plate to be contacted with the top surface of porous stone at the upper part of the sample, and the sample is not affected by the overlying load. The measurement method comprises steps as follows: static-pressure sample preparation is carried out according to geotechnical test rules of roads, the sample is soaked in water in constant volume, vertical expansive force is measured, a direct shear test is performed with the vertical expansive force as the overlying load, and effective cohesion and effective internal friction angle of experimental expansive soil are obtained. The test device and the method have the advantages of being simple, convenient to operate, reasonable in structure and economical and guarantee consistent micro-structure before sample shear under the action ofdifferent overlying loads in the direct shear test, so that the problem that shear strength parameters of expansive soil cannot be accurately obtained for a long time in the field is solved, and accurate design parameters are provided for civil engineering structure design of the expansive soil.

The invention discloses a concrete T-shaped beam post-fire damage evaluation method, and relates to the technical field of post-fire building damage grade evaluation methods, and the concrete T-shapedbeam post-fire damage evaluation method comprises the following steps: 1, determining parameter indexes; 2, according to the determined parameter indexes, obtaining the comprehensive closeness of each evaluation object by utilizing a plurality of evaluation methods, wherein the multiple evaluation methods comprise an improved analytic hierarchy process, an entropy weight method, an approximate ideal solution, a deviation quadratic sum optimization method, a grey relational degree method and an orthogonal projection method; and 3, finally determining the damage grade of the building accordingto the comprehensive relative closeness of each evaluation object. According to the concrete T-shaped beam post-fire damage evaluation method, a set of comprehensive evaluation system of a grey relational degree correction improved approximate ideal solution (MTOPSIS-GRA) method based on the optimal combination weight is established, so that the evaluation result is more objective, and the technical defects of the evaluation method in the prior art are avoided, and the concrete T-shaped beam post-fire damage evaluation method is suitable for engineering application and popularization.

The invention relates to an image repairing method based on sample structure consistency. The image repairing method comprises the following steps: determining an image defect area boundary (the symbol of which is shown in the description) and calculating a priority function value of each pixel point p on the image defect area boundary; selecting the pixel point (p') with a maximum priority value P(p) from the pixel points on the image defect area boundary and then finding matching blocks which is the most similar to a target block (psi p') in a known image area; finding the optimal matching block (psi q') and then enabling unknown pixel points in the target block (psi p') to be filled with corresponding pixel values in the optimal matching block (psi q'); updating the confidence of the filled pixel points; repeating the above steps till the defect area is completely filled.

The invention discloses a text retrieval method and a relevant device. The method comprises: Input primary text data; perform self-adapting mapping process with a method of descent for the primary text data; carry forth retrieval text data similar to the data after the self-adapting mapping process with the method of descent; output the retrieval text data. The invention execution example can fulfill effective compression for original data after decency for higher-dimensional text data; the invention is characterized in rather low time complexity, adapting to magnanimity data, and effectively maintaining similarity between all vectors of the text data. Utilization of the invention execution example method or the device can fulfill quick response to such requests as network text inquiry, search, retrieval and etc at rather low arithmetic cost, so as to resolve the problems in prior network text retrievals of high time complexity, large space consumption and rather low accuracy.

The invention comprises a multi-nozzle rocket jet test system. The system comprises a simulated rocket body, a test platform and a water spraying device used for reducing noise and protecting the testplatform; the outer contour dimension of the simulated rocket body and the outer contour dimension of the test platform are designed in proportion according to the actual structure size or are subjected to linear scaling, the total water spraying flow of the water spraying device is k<2> times of the total actual water spraying flow, and the water spraying speed of the water spraying device is consistent with that of an actual product; and the simulated rocket body is fixedly installed on the test platform according to a preset test height, and produces fuel gas during a test, temperature, pressure, heat flow and noise in the test process are collected through sensors arranged on the simulated rocket body and the test platform, and the k is the ratio between the outer contour size of thesimulated rocket body and the outer contour size of the test platform to the actual structure size of the same.

The invention relates to a face super-resolution method based on a supervised pixel-by-pixel generative adversarial network. The method comprises the following steps: firstly, reading an original facepicture data set; then carrying out preprocessing work such as data cutting and cleaning; thirdly, reading the high-resolution face image for bicubic interpolation down-sampling to obtain a high-resolution face image-low-resolution face image pair; fourthly, inputting the low-resolution face image into a generator network to generate a super-resolution face image; and fifthly, respectively inputting the high-resolution face image and the super-resolution image into a pixel-by-pixel discriminator network, calculating a supervised pixel-by-pixel adversarial loss function by using a pixel-by-pixel discrimination matrix output by the pixel-by-pixel discriminator network, and training a generator network by using error back propagation.

The invention belongs to the technical field of food safety immunity detection, and specifically discloses a semi-antigen, artificial antigen, and antibody aiming directly at histamine, a preparation method and applications thereof. The molecular structure of the semi-antigen is represented by the formula (I), which is shown in the description, wherein the n=0-6. The semi-antigen can combine with a carrier protein to obtain the artificial antigen, and the antibody obtained from animals immunized by the artificial antigen can directly recognize histamine, while no obvious cross reaction exists between the antibody and other substances with a similar structure as the histamine. Compared to the indirect derivation method, which has been reported before, the histamine immunoanalysis method, which is built on the basis of the antibody, saves the derived sample pretreatment step, so the organic reagent consumption and analysis time are saved, and the detection cost is reduce. The titer of the polyclone antibody is 1:64000, the semi-inhibition concentration is 56.3 mg / L, and the lowest detection limit of histamine can reach 25.8 mg / L lower than the prescribed limit standards of every country.