40results about How to "Help digging" patented technology

The invention discloses a single image rain line removal method based on the deep convolutional neural network. The method comprises steps that firstly, a rain image is decomposed into a low frequencyreference layer and a high frequency detail layer through utilizing guided filtering; secondly, a target function is modified according to the knowledge of the image processing domain, the high frequency detail layer of the rain image is inputted into the designed deep learning network architecture, and mapping between the high frequency detail layer of the rain image and the high frequency detail layer of a sharp image corresponding to the rain image is learned; and lastly, the rain-removed high frequency detail layer outputted by the network and the low frequency reference layer of the rainimage are superposed to obtain the rain-removed sharp image. The method is advantaged in that the detail part of the rain-removed image is remained when rain lines of a single image are removed, andthe image sharpness degree is greatly improved.

The invention discloses a multi-document subject discovery method based on two-layer clustering. The multi-document subject discovery method comprises the following steps: S1 using a plurality of documents as input, pretreating each document, i.e. the documents are broken up into clauses, and the clauses are broken up into words, so as to obtain a noun group and a verb group in a multi-document group, and performing emantic disambiguation processing on polysemes in the noun group and the verb group; S2 respectively performing word clustering analysis on the noun group and the verb group which are output in the step S1 according to word similarity by adopting an improved OPTICS algorithm, extracting semantic concepts, and establishing vector space models on the clauses according to the semantic concepts; S3 performing clustering analysis on the clauses by using an improved K-medoid algorithm, so as to obtain a subject. Inner semantic relations between words are extracted by the multi-document subject discovery method, and the problem of non-orthogonality among feature items when feature vectors of the clauses are established is solved.

The embodiment of the invention discloses a method and device for evaluating the uniformity of cascaded battery packs. According to the method and the device, the disadvantages that the difference of intrinsic parameters (such as open-circuit voltage, available capacity and internal resistance) of single batteries in the battery packs is difficult to be truly described, evaluation indexes are not comprehensive enough, and the applicability is relatively poor in an existing uniformity online evaluation technique are solved. A distance similarity index calculation method based on a dynamic time warping strategy is developed based on a measured voltage curve of the single batteries, influence factors which cause the voltage difference of the single batteries due to different SOCs can be eliminated in a self-adaption manner, the evaluation of operational characteristic difference of the single batteries in charging-discharging processes is realized, and compared with an existing uniformity extraction method based on voltage difference of each discrete moment, intrinsic information of the batteries is assisted to be excavated; and on this basis, an abnormal battery statistics recognition method based on a stepwise iteration strategy is developed, the uniformity evolution of the cascaded battery packs is realized, and the effective technical support is provided for operation maintenance.

The invention discloses a rebate platform and method based on graded accumulated points. The rebate platform based on graded accumulated points includes a database storage rebate qualification standard, business rebate rules and consumer graded rebate rules. A rebate server comprises a business turnover acquiring module, a consumer consumption sum acquiring module, a rebate qualification determining module and a rebate module, wherein the rebate module is used for calculating the sum of the surrendered profits that the business should pay for the platform according to the business rebate rules and the business turnover, and gives the rebate to the consumers according to the sum of the surrendered profits and the consumer graded rebate rules. For the rebate platform and method based on graded accumulated points, the consumers can get profits and the business surrendered profits; and by means of integration and data operation through the platform, the multi-win aim that the consumers can get full rebate of the sum of consumption can be achieved. The platform counts the daily sales of all the cooperated business and collects and takes the surrendered profits of the business, and then determines the sum of consumption and conditions of the consumers, and then calculates and accumulates sum of consumption of every time for the consumers, and then by combination of the surrendered profits of the business, and according to the consumption proportion, can perform full rebate for the consumers day by day.

The invention discloses a production process of a honeycombed SCR (selective catalytic reduction) catalyst with pyrophyllite as a carrier, belonging to the fields of nitric oxide after-treatment and purification and environment-friendly catalytic materials. The catalyst has three layers of microstructures, wherein pyrophyllite forms a carrier layer; tungsten-doped titanium dioxide (TiO2-WO3) forms a transition layer; and vanadium pentoxide (V2O5) forms an active layer. The catalyst has the beneficial effects that the negative effects of trace elements Na and K in pyrophyllite on the activity of the honeycombed SCR catalyst can be avoided by the design of the three layers of microstructures; pyrophyllite has the characteristics of low moisture expansion and high heat stability; as the carrier, pyrophyllite is favorable for improving the yield in the catalyst drying process and is favorable for improving the mechanical strength of the catalyst, and the cost of the catalyst can be also reduced by reasonably adding pyrophyllite; and the catalyst has high lateral and longitudinal mechanical strength, has denitration efficiency of more than 89% at a temperature of 270-500 DEG C, and has denitration efficiency of more than 94% at a temperature of 330-430 DEG C.

The invention provides a potential talent determination method and an electronic device. The potential talent determination method comprises the following steps: obtaining the evaluation scores of n basic abilities of the object to be tested for multiple times in a database and presetting evaluation coefficients corresponding to the month-age segment, wherein n is greater than or equal to 1; obtaining the evaluation scores of n basic abilities of the object to be tested for multiple times in a database; calculating a potential value An of the target basic ability in each month age segment based on the evaluation score of the multiple evaluation, the evaluation coefficient and the preset potential calculation formula; based on the potential of the target 's underlying capabilities, the preset potential coefficient corresponding to the month-age segment and the preset potential talent calculation formula calculate the potential talent score of the target basic ability of the object to betested, thereby improving the accuracy of the evaluation of the basic ability of the object to be tested, and assisting parents in digging out the technical effect of the real talent hidden in the child.

The invention relates to a mining method for APP software user comments reflecting user intentions, and belongs to the field of APP software user intention analysis. The method comprises the followingsteps: firstly, defining an evaluation object and an extraction rule of an evaluation viewpoint, initializing a comment seed on the basis of extracting the evaluation object and the evaluation viewpoint of a user comment, and mining the user comment which is the same as or similar to the comment seed and embodies the same user intention category by applying the comment seed; then, user comments reflecting user intentions in the user comments failed in matching with the comment seeds are mined through willingness verbs, software brief introduction and sentiment analysis, and a candidate comment mode library is constructed; and finally, extracting representative new comment seeds for the candidate comment mode library constructed in each cyclic mining process for the next cyclic mining to realize the cyclic mining of the user comments embodying the user intention until the mining effect of the user comments embodying the user intention is convergent.

The invention relates to a photovoltaic output power ultra-short-term local emotion reconstruction neural network prediction method, and the method employs a singular spectrum analysis method to reduce the noise of original photovoltaic power, and can reduce the complexity of a photovoltaic power time sequence. Implicit fluctuation information mining of a photovoltaic power time sequence through achaotic nonlinear dynamics method is facilitated, meteorological data does not need to be acquired in advance in actual engineering, and the problem of model complexity caused by improper selection of input parameters and the problem of error accumulation caused by inaccurate prediction of the meteorological data can be avoided; the built local emotion reconstruction neural network does not needto determine the number of hidden layer nodes, modeling is simple, and the prediction cost is reduced; the established local emotion reconstruction neural network determines expansion signals and emotion parameters according to photovoltaic power chaotic attractors, so that the prediction model more pays attention to tracking each input mode and selects the most useful input mode information, themapping relationship is more accurate, the prediction accuracy is higher under different weather conditions, and the adaptability is higher.

The invention belongs to the field of deep learning analysis, and discloses a method for scoring the curing probability of a potential treatment scheme based on patient detection information. The method comprises the following steps: (S1) collecting training data; (S2) establishing a logistic regression model and training the logistic regression model; (S3) respectively establishing a unet model and a three-dimensional convolutional neural network model and training the unet model and the three-dimensional convolutional neural network model; (S4) complementing the training data to obtain a survival condition score, corresponding to the training data, of any patient in each day in the treatment process; (S5) establishing a long short-term memory (LSTM) network model and training; and (S6) randomly selecting the potential treatment scheme, and predicting to obtain a survival condition score of the (N + 2) day, thereby completing scoring of the potential treatment scheme. The overall process design and the like of the method and the system are improved, the condition of a patient can be specifically quantified from the angle of big data analysis, and the curing probability of a potential treatment scheme is scored.

The invention discloses a self-mixing type anaerobic ammonia oxidation baffle plate reactor which comprises a water bath heating system and a reactor body, wherein the water bath heating system is arranged on the reactor body and comprises a heater and a water bath heating layer; the water bath heating layer wraps the two side walls and bottom of the reactor body, and the heater is connected with the water bath heating layer; the reactor body comprises a water distribution area, a reaction area and a precipitation area; the water distribution area comprises a water inlet formed at the bottom of the reactor body; the reaction area is arranged above the water distribution area, multiple baffle plates are vertically arranged in the reaction area, and the reactor main body is divided into at least four cells in series connection by the baffle plates. By adopting the self-mixing type anaerobic ammonia oxidation baffle plate reactor disclosed by the invention, the wastewater mixing can be promoted, the problem of non-uniform microbial activity and number caused by non-uniform matrix distribution of the cells is effectively solved, and the overall reaction potential of the reactor is improved.

The invention discloses a channel, group and group user activeness determination method, system and terminal, and relates to the technical field of network data processing. Acquiring data of channels, groups or group users in observation time, acquiring probability distribution of the interaction amount in the observation time according to the data, and performing data preprocessing on the acquired probability distribution data of the interaction amount; and obtaining user activeness interval distribution data, and performing activeness discrimination based on the obtained user activeness interval distribution data. According to the method and the device, the activeness of the channel, the group and the group user can be quickly and accurately judged, the relationship between the group and the group user can be mined, and the parts calling the group activeness can be conveniently acquired from the group.

The invention belongs to the field of image analysis, and discloses an automatic classification method and system for pathological diagrams of cell and regional hierarchies, and the method comprises the following steps: (1) cutting an initial pathological diagram into small diagrams, and drawing a cell mask diagram; (2) carrying out image segmentation on the small image to obtain cell coordinates; (3) cutting the initial pathological map into single cell small maps, and judging cell types; (4) cutting the initial pathological diagram into small diagrams, and judging whether the small diagrams belong to a cancerous region and a para-cancerous region or not; (5) cutting the initial pathological diagram into small diagrams, and respectively judging whether the small diagrams belong to a tumor tissue parenchyma region, a mesenchymal region, a protein keratinization region and a vessel region or not; (6) integrating the scores, and determining the cell type and the belonging region thereof. According to the method, the overall process processing of the method is improved, so that the defect that the analysis of the pathological diagram stays in block type analysis in the prior art can be overcome, the information of the pathological diagram is deeply mined, and therefore, the automatic classification of the pathological diagram is realized.

The invention provides an aspect-level sentiment analysis method and device and a computer readable storage medium, and the method comprises the steps: providing a target text to an input layer of a convolutional neural network, and obtaining a corresponding input feature matrix; providing the input feature matrix to a first convolutional layer to obtain a corresponding first feature map, and processing the first feature map by using a first activation function to obtain a first feature vector, wherein the first feature map reflects emotional feature information of the target text; processing the first feature map according to the aspect feature information of the aspect to be analyzed to obtain a second feature map, and processing the second feature map by adopting a second activation function to obtain a second feature vector, wherein the second feature map reflects association information of the target text and the aspect to be analyzed; and judging the emotional polarity of the target text corresponding to the to-be-analyzed aspect according to the first feature vector and the second feature vector. The method enables the model to be simple in structure and low in operand.

The invention discloses a service processing method, device and system, and at least solves the problems that in the prior art, a mode of obtaining service data from a service platform is small in data load and is limited by compatibility of an operating system and terminal equipment. The service processing system comprises a BIP (Bearer Independent Protocol) gateway platform, a service platform and an SIM (Subscriber Identity Module) card, wherein the service platform is used for generating an APDU (Application Protocol Data Unit) instruction based on service data required by the SIM card, and sending the APDU instruction to the BIP gateway platform through a BIP channel between the service platform and the BIP gateway platform; the BIP gateway platform is used for forwarding the APDU instruction to the SIM card through a TCP (Transmission Control Protocol) channel between the BIP gateway platform and the SIM card; and the SIM card is used for analyzing the APDU instruction to obtain the service data, and executing corresponding service processing based on the service data.

The invention provides an overall microcirculation function evaluation of lung tissue and visual presentation method, which comprises the following steps: capturing lung tissue microcirculation oxygen parameters and microcirculation hemodynamic parameters, cleaning lung tissue microcirculation function data, and finally, carrying out general microcirculation framework and wavelet analysis method to visually present the overall microcirculation function of the lung tissue. Compared with the prior art, the whole microcirculation function of the lung tissue is evaluated more comprehensively and completely from the two aspects of lung tissue microcirculation hemodynamics and microcirculation oxygen; meanwhile, null value data and abnormal data from instrument firmware and optical components are eliminated through a microcirculation function data cleaning module; finally, the lung tissue microcirculation function data is visually presented, the lung tissue microcirculation function state can be deeply analyzed and excavated, and the scientificity of lung tissue microcirculation function characteristic display is improved.