73results about How to "Alleviate the lack" patented technology

The disclosure provides a method, an electronic apparatus, and a computer readable medium of constructing a classifier for disease detection. The method includes the following steps. A codebook of representative features is constructed based on a plurality of disease-irrelevant data. Transfer-learned disease features are extracted from disease-relevant bio-signals according to the codebook without any medical domain knowledge, where both the disease-irrelevant data and the disease-relevant bio-signals are time-series data. Supervised learning is performed based on the transfer-learned disease features to train the classifier for disease detection.

The invention relates to a method of qualitatively evaluating a digital audio signal. It calculates a quality indicator consisting of a vector associated with each time window in real time, in continuous time, and in successive time windows. For example, the generation of a quality indicator vector calculates, for a reference audio signal and for an audio signal to be evaluated, the spectral power density of the audio signal, the coefficients of a prediction filter, using an autoregressive method, a temporal activity of the signal or the minimum value of the spectrum in successive blocks of the signal. To evaluate the deterioration of the audio signal, the method may calculate a distance between the vectors of the reference audio signal and the audio signal to be evaluated associated with each time window.

The invention provides a semi-supervised text classification model training method and system, a text classification method and system, equipment and a medium. The training method comprises the stepsof acquiring an initial sample set; enhancing the unlabeled sample to obtain a data enhanced sample; inputting the unlabeled sample and the data enhancement sample into a text classification model toobtain an embedded vector and a prediction probability belonging to each classification label; for each unlabeled sample, obtaining an embedded vector mean value of the unlabeled sample and the corresponding data enhancement sample as a new sample; for each unlabeled sample, obtaining a label estimation result of a new sample after sharpening the prediction probability mean value of the unlabeledsample and the corresponding data enhancement sample belonging to each classification label; verifying whether the new sample is credible or not, and if yes, marking the new sample as a credible new sample; constructing a confidence sample set according to the labeled sample and the trusted new sample, and expanding to obtain a target sample set; and training the text classification model according to the target sample set. According to the method, the text classification accuracy is improved under the condition of lack of annotation samples.

This invention relates to a compound type polymer dielectric stuff and its preparation method, belong to electrochemistry and new energy stuff region. It composed by copolymer of fluoro ethylene and hexafluoro-propylene, lithium salts and plastic crystal stuff. Processing steps: take copolymer, lithium salts and plastic crystal stuff, use solvent to dissolve, whip go as far as uniformity, then pour above solution to teflon tooting, evaporate solvent, finally vacuum drying. While this compound type electrolyte stuff in rather wide temperature range and all components are solid-state, the ionic conductivity can reach using standard of lithium Secondary Battery. It overcomes shortcoming of traditional gel-type polymer dielectric that liquid electrolyte easily leak as well as plastic crystal electrolyte unease form film, and also combine the merit that polymer dielectric easily form film as well as plastic crystal electrolyte has high room temperature ionic conductivity.

The invention relates to meat duck feed and a processing method thereof. The feed comprises the following ingredients in percentage by weight: 42.83 percent of maize, 19 percent of wheat shorts, 13 percent of beam dregs, 8 percent of cotton dregs, 5 percent of rape seed cake, 3 percent of maize germ cake, 2.0 percent of maize protein powder, 0.8 percent of lard, 1.3 percent of mountain flour, 1.5 percent of calcium hydrophosphate, 1.5 percent of fish meal, 0.21 percent of common salt, 0.58 percent of lysine, 0.15 percent of methionine, 0.13 percent of threonine, 0.2 percent of composite bio-enzyme preparations and 1.0 percent of premix. The processing method comprises the following steps of: crushing and / or grinding all ingredients in the feed; then, adding various raw materials into a stirring machine to be stirred; and finally, packing the stirred feed into bags. The conversion rate, the quality stability and the dreg type protein consumption of the feed are improved through the use of additives, so the cost of the feed is reduced, and the growth expression of meat duck can be maintained or improved.

The invention relates to a piglet low-protein feed, in particular to a piglet low-protein feed with lower protein content. The piglet low-protein feed comprises the following compositions in portions by weight: 30-40 portions of corn, 25-35 portions of puffed corn, 2-4 portions of imported fish meal, 10-15 portions of dehulled soyabean pulp, 4-8 portions of enmolysis soyabean pulp, 2-6 portions of whey powder, 1-3 portions of soyabean oil, 1-2 portions of calcium hydrophosphate, 0.1-1 portions of mountain flour, 0.2-0.6 portion of lysine, 0.1-0.4 portion of methionine, 0.05-0.2 portion of threonine, 0.01-0.1 portion of tryptophane, and 2 portions of premixing compound. The invention reduce the use of protein raw materials during the preparation of the feed, and supplements crystal amino acid, thereby balancing the level of the amino acid in the feed, relieving the problem of lacking the protein raw materials on the premise of meeting the demand of piglets to a certain extent, loweringthe feed cost, reducing the environmental pollution of a farm, and reducing the diarrhea of the piglets.

The invention discloses a mobile-phone-type direct ophthalmofundoscope. The mobile-phone-type direct ophthalmofundoscope comprises a shell, wherein the lower end of the shell is a handheld end; the upper end of the shell is a video imaging end; a lighting module, an observation module, a fundus imaging module and an image transmission module are arranged in the shell; the lighting module comprises a light source, a light intensity adjusting device, a diaphragm disc, a condenser lens and a reflecting mirror; the condenser lens and the diaphragm disc are arranged in a light path between the light source and the reflecting mirror sequentially; the light source is arranged at the handheld end; the reflecting mirror is arranged at the video imaging end; and an observation system module, the fundus imaging module and an imaging display module are arranged at the video imaging end. The invention provides the mobile-phone-type direct ophthalmofundoscope, which is convenient to carry by an examiner and can photograph the fundus image and transmit the fundus image in a wireless mode.

The invention provides a text classification method based on expanded label samples. Firstly, a real sample data set containing labeled text samples and unlabeled text samples is collected; then high-dependability samples are found through a clustering method KFCM to obtain the expanded label samples; then a square loss function is utilized to formulate a uniform classification objective functionfor labeled, unlabeled and extended sample data, parameters of regularization parameters, a kernel function and the like therein are set, and a text classification function is obtained by learning; and finally, to-be-classified text data are input, the text classification function is utilized for classification, and a class of text is obtained. The invention also provides a text classification system. Compared with error rates of other classical classification algorithms and related algorithms on test sets, error rates of the method are significantly improved. The method solves the problem oflow classification precision of the prior art on text with fewer and inaccurate label samples, and has best mutual information and class variables.

The invention discloses an SQL (Structured Query Language) conversion method and system based on language model coding and multi-task decoding. The method comprises the steps: combining a language model in combination with a field where a data set is located to carry out pre-training, and improving the feature extraction capability in the field; sequentially expanding the query database according to table names and column names, converting a two-dimensional table into a one-dimensional text sequence, and splicing the one-dimensional text sequence into an input sequence X in combination with user questions; inputting the sequence X into a pre-training language model, and outputting a coding result; a multi-task decoder composed of nine different neural networks is utilized to decode and restore the SQL fragments, and cross entropy loss is calculated; different weights are set for loss values of different neural networks, the sum is finally calculated as the total loss of the model, a gradient descent algorithm is utilized to optimize an objective function, and model training parameters are updated; after training is completed. Model parameters are stored, and a corresponding SQL sequence is automatically generated according to the user problem and the target database.

The disclosure provides a method, an electronic apparatus, and a computer readable medium of constructing a classifier for skin-infection detection. The method includes the following steps. A codebook of representative features is constructed based on a plurality of target-disease-irrelevant images. Transfer-learned disease features are extracted from target-disease images according to the codebook without any medical domain knowledge, where the target-disease images are captured by at least one image capturing device. Supervised learning is performed based on the transfer-learned target-disease features to train the classifier for skin-infection detection.

The invention relates to an intelligent multi-function guiding robot and a control method thereof comprising a robot body and a handheld control unit. The robot body comprises a head portion, a support portion, and a walking mechanism; the head portion is provided with a scene recognition device, an infrared sensor module, a display screen, and a loudspeaker; a first processing unit is arranged inthe head portion, and the scene recognition device, the infrared sensor module, the display screen, and the loudspeaker are all electrically connected with the first processing unit; the support portion is connected between the head portion and the walking mechanism; and the handheld control unit is provided with a two-dimensional code and a control button, and the scene recognition device activates the robot to starting up and communication by the scanning the two-dimensional code. The intelligent multi-function guiding robot and the control method thereof solves the problems that at presentprofessional explainers have difficulty in making appointment, personnel are scarce, and explain quality is low, more tourists are attracted to visit, cultural transmission function such as technology exhibition halls and heritage museums is further played, and service quality and management level of sites such as the technology exhibition halls and the heritage museums are enhanced.

The invention discloses a method for producing white carbon black by using high alumina fly ash and a system thereof, relating to the technical field of material and chemistry. The invention comprises a fly ash slurry blending system, a fly ash pre-desiliconization system, a white carbon black preparation system and a caustification system. the structure is characterized in that the sleeve heater is divided into two sections; the outlet passage at the lower part of a heat preservation stay pot is divided into two paths; the solution outlet of a tube type falling film evaporator is connected with one inlet of a circulating mother liquor preparation tank through an evaporating solution pump, and one of the other two inlets of the circulating mother liquor preparation tank is alkali supplementing passage and the other thereof is connected with the outlet of a caustification tank through a caustification solution pump; the outlet passage of the circulating mother liquor preparation tank is connected with the inlet of a fly ash slurry preparation tank through a circulating mother liquor pump, so that the whole system forms a closed circulation. The invention is simple in process, low in production cost, low in energy consumption, and stable in performance of the products, meanwhile the fly ash of which white carbon black is extracted out can be used for preparing aluminium oxide as an important aluminium resource.

The invention discloses an automatic segmentation method of a dangerous organ based on a convolution neural network, belonging to the technical field of image processing. The method comprises the following steps: S1, acquiring patient CT image data and corresponding labeling data; S2, preprocessing the CT graphic data and the corresponding labeling data; 3, establish a 3D convolution neural network model, inputting data block, and obtaining a prediction result image output by that model; S4: optimizing the prediction result image output by the 3D convolution neural network model. The present invention is only based on CT image data, has small difficulty in obtaining original data and wide application range, and can realize automatic segmentation of dangerous organs in CT image, and the segmentation process does not require manual interference, effectively improves segmentation efficiency and segmentation result precision, increases post-processing operation, and further optimizes the segmentation result.

The disclosure provides a method, an electronic apparatus, and a computer readable medium of constructing a classifier for disease detection. The method includes the following steps. A codebook of representative features is constructed based on a plurality of disease-irrelevant data. Transfer-learned disease features are extracted from disease-relevant bio-signals according to the codebook without any medical domain knowledge, where both the disease-irrelevant data and the disease-relevant bio-signals are time-series data. Supervised learning is performed based on the transfer-learned disease features to train the classifier for disease detection.

The invention relate to a pregnant sow low protein daily feed formula, belonging to the animal feed field. The pregnant sow low protein daily feed formula provided in the invention can reduce production cost, maintain or increase the growth performance of the pregnant sow, to solve the problems existing in the current technologies. The invention increases the raw material utilization rate of pregnant sow, meets the needs of pregnant sow by reducing the soybean meal content in daily feed and supplementing some artificial synthesized essential amino acids, while maintains the production performance constant, and improves feed conversion rate, livestock and poultry nitrogen deposition and nitrogen discharge, and heat stress capability.

The invention relates to the field of livestock breeding, and specifically discloses a novel application of anthocephalus chinensis as a feed. Anthocephalus chinensis comprises approximately 17% of crude protein, which is equivalent to that of alfalfa. According to the invention, as a result of researches, no negative influence is caused to animal production and slaughter performances if a certain proportion of anthocephalus chinensis leaf instead of alfalfa powder is added into a feed. Therefore, anthocephalus chinensis can be used as a novel feed resource for replacing alfalfa. Therefore, the lack of plant protein forage of our nation can be relieved.

An image capturing apparatus comprises: an image-capturing unit acquiring RAW image data indicating a pre-development image from an image signal; recording unit encoding the RAW image data by a predetermined encoding method and recording the encoded RAW image data in a medium; a developing unit reading out the encoded RAW image data and developing the read encoded RAW image data; and a control unit comparing an amount of the encoded RAW image data that was recorded in a shooting time period with an amount of the encoded RAW image data that was read out from the medium so as to be used in development, and controlling an encoding method by which the RAW image data is encoded according to a comparison result.

The invention relates to a method of qualitatively evaluating a digital audio signal. It calculates a quality indicator consisting of a vector associated with each time window in real time, in continuous time, and in successive time windows. For example, the generation of a quality indicator vector calculates, for a reference audio signal and for an audio signal to be evaluated, the spectral power density of the audio signal, the coefficients of a prediction filter, using an autoregressive method, a temporal activity of the signal or the minimum value of the spectrum in successive blocks of the signal. To evaluate the deterioration of the audio signal, the method may calculate a distance between the vectors of the reference audio signal and the audio signal to be evaluated associated with each time window.

The invention discloses a low-nitrogen-emission daily ration for growing pigs. The low-nitrogen-emission daily ration contains protein, lysine, tryptophan, threonine and sulfur-containing amino acid, wherein the net energy content of the low-nitrogen-emission daily ration is 2.30 to 2.45 Mcal / kg, and the content of the protein is 11.5 to 15.5%; the content of the lysine capable of being digested by standard ileums in the low-nitrogen-emission daily ration is 0.9 to 1.02%, the ratio of the tryptophan capable of being digested by the standard ileums to the lysine capable of being digested by the standard ileums is not lower than 0.23, the ratio of the threonine capable of being digested by the standard ileums to the lysine capable of being digested by the standard ileums is not lower than 0.66, and the ratio of the sulfur-containing amino acid capable of being digested by the standard ileums to the lysine capable of being digested by the standard ileums is not lower than 0.61. According to the low-nitrogen-emission daily ration disclosed by the invention, by decreasing the crude protein level in a daily ration prescription in Feeding Standard of Swine (2004) and supplementing synthetic amino acids, the absorptivity of the amino acids in animals is increased, the feed cost is lowered, and the economic benefits of swine industry is enhanced; the decrease of nitrogen emission in environment is beneficial to alleviation of the problem of protein feed insufficiency to a certain extent; and the weight gain of growing pigs and the feed conversion rate are not influenced.

The invention relates to tiny algae cultivating field, especially the high density cultivating method of Tama Alexander algae. In the initial stages, static laying and mildness whisking would be adopted to make the carbon source and other alimentation in the culture medium proliferate for a period of time. Taking aerate after the cell density reaching a certain extent, thus, injure of cell would be avoid. The gas could have a certain ratio CO2 to supply carbon source that the algae cell needs for growing. And a certain PH value would be sustained. At the initial stages, suitable N, P consistence would be supplied to make the cell grow fast. And at the late stages, N, P should be added to relief the lack of N, P, and achieve high cell density.

Flaws in dependencies of deployed applications are identified. In one embodiment, a list of dependencies used by a deployed application that is deployed on the deployment platform is obtained. Each dependency of the list of dependencies is mapped with a flaws database, wherein the flaws database comprising an indication of known flaws for different dependencies and different versions thereof. Based on such mapping, one or more flaws in the deployed application are determined. The determination is performed externally to the deployment platform and without executing a monitoring process thereon.

The invention discloses a therapeutic antibody for feline distemper. The therapeutic antibody for feline distemper is prepared from an FPV inactivated vaccine and an FPV subunit vaccine as immunogensby horse immunization. Therefore, a corresponding preparation method is established; according to the method, two different types of FPV immunogens are creatively inoculated to a healthy horse according to a certain program, the FPV inactivated vaccine is taken as a basic immunogen, the FPV subunit vaccine is taken as an enhanced immunogen, hyperimmune blood is collected after a high-potency antibody is produced, and high-activity immune globulin F(ab')2 is prepared by IgG purification and Fc' segment excision. With the application of the method, a large amount of horse-derived anti-FPV hyperimmune serum can be prepared, and the high-potency F(ab')2 antibody is extracted and purified. A test indicates that the prepared therapeutic antibody for feline distemper not only has higher potency,but also has exact feline distemper preventing and treatment effects, can be used for emergency prevention and treatment of feline distemper and has important practical significance and broad application prospect.

The invention discloses a cutting propagation method of tinospora capillipes gagnep, aiming to solve problems of large input and laborious operation of cutting propagation methods in the prior art, which is not conductive to large-scale popularization and use. The cutting propagation method includes the steps of 1), cutting preparation, namely selecting woody and semi-woody tinospora capillipes gagnep vines, cutting the vines by 25-35cm as cuttings for standby use; 2), cutting treatment and cutting, namely cutting a part of each blade on the cuttings, soaking the whole cuttings with 100-300mg / L of an indole butyric acid solution, and then burying the cuttings in a cutting matrix for culturing. The method is simple, low in input, convenient in operation and high in the survival rate of thecuttings, capable of rapidly breeding tinospora capillipes gagnep, and beneficial to large-area popularization and use.

The invention belongs to the field of tobaccos, and particularly relates to a method for slitting and using the tobacco leaves, which comprises the following steps of: equally dividing the tobacco leaves to be slit into three to seven sections along the main vein direction to obtain tobacco leaf slitting sections, wherein the tobacco leaves to be slit are selected from one or more grades of tobacco leaves of B2F, C3F and X2F; obtaining the total nicotine and total sugar content of the tobacco leaf slitting sections and grade tobacco leaves, wherein the grade tobacco leaves are selected from one or more grades of tobacco leaves of B1F, B2F, C2F, C3F, C4F, X2F, B1K, B2K, C3V, X2V, B2V, S1 and CX1K; taking the total nicotine content and the total sugar content as indexes to carry out clustering analysis on tobacco leaf slitting sections and grade tobacco leaves; using the tobacco leaf slitting sections of the same class to replace tobacco leaves of the same grade. The invention further relates to a device for slitting tobacco leaves and using the tobacco leaves and a tobacco product. The method relieves the problem of lack of grade tobacco leaf raw materials, and is simple in slittingoperation and suitable for industrial popularization.

According to the present invention, there is provided a variable tension fastener for attaching, and tensioning, a cover over a vehicle. The two piece fastening system includes a first member for attachment to a peripheral portion of a vehicle and a second member for attachment to a cover, one of the members comprising a loop, and the other of the members comprising a hinged lever pivotal about a pivot and including a male member for extending through the loop and having a series of transverse recesses spaced sequentially from a pivot for selective engagement with the loop, where pivoting of the lever about the pivot, to a closed position, tensions the cover and attaches the cover to the vehicle.

The invention discloses a forage grass planting method based on mixed cropping of oat and common vetch. The method comprises the following steps of fine soil preparation before sowing and combined application of N fertilizer, P fertilizer, K fertilizer and organic fertilizer according to the soil fertility; variety selection, wherein the seed quality reaches the national level 3 or above, oat for the areas at the altitude of 3,000 m or above is selected from early maturing varieties such as Qingyin No.1, common vetch for the areas at the altitude of 3,000 m or above is selected from early maturing varieties such as Lanjian No.1, oat for the areas at the altitude of 3000 m or below is selected from the varieties such as Qinghai sweet oat, and common vetch for the areas at the altitude of 3000 m or below is selected from the varieties such as Ximu 324; determination of the sowing amount, wherein the sowing amount of the oat and common vetch is determined according to respective seedling conservation numbers, the seedling conservation number of the oat is 4,500,000-5,250,000 plants per hectare, and the seedling conservation number of the common vetch is 3,000,000-3,750,000 plants per hectare. The method has the advantages that the fertilizer application amount under different soil fertility levels is stipulated; the selection of planting varieties for mixed sowing forage grass land in different areas is stipulated; a method for planting in the mixed sowing forage grass land is stipulated; land resources can be comprehensively utilized, natural resources such as light, heat, fertilizer and water can be utilized intensively, the cultivation intensity and the planting cost can be reduced, fertilizer application is reduced, and the land resources can be protected.

The invention relates to a catalytic cracking adsorption-stabilization system with gasoline cutting and a technique for increasing the content of propylene and reducing the content of olefins. The catalytic cracking adsorption-stabilization system and the technique are as follows: a gasoline cutting tower is connected to the rear of a double-tower flow stabilizer; a stream of gasoline is stabilized at the bottom of the stabilizer and enters into the central section of the increased gasoline cutting tower at a temperature of between 105 and 115 DEG C; after the gasoline is cut by the gasoline cutting tower, a part of light gasoline on the top of the tower enters into a reaction-regeneration system for remixing; heavy gasoline at the bottom of the tower is mixed with the light gasoline at a temperature of between 35 and 45 DEG C after a heating desorption tower 10 is fed; and mixed streams pass through a water cooler and are outputted from the device when the temperature is reduced to be lower than 40 DEG C. By utilization of the flow of the catalytic cracking adsorption-stabilization system and the technique, the light gasoline after cutting is remixed, thereby the content of the propylene can be increased, thereby being favorable for solving the problem of insufficient propylene products in the petrochemical industry of China. Moreover, because the gasoline cutting is clear, the content of olefins in the gasoline is greatly reduced and the quality of the gasoline products is obvious improved. The energy-saving technique can reduce the energy consumption during the stable absorption process and reduce the operating cost.

The present invention provides a method for the diagnosis of disorders caused by foetal alcohol syndrome, said method comprising the assaying of PLGF (placental growth factor).