Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

38results about How to "Meet expectations" patented technology

Multi-mode RAG knowledge question-answering method and device applied to vertical field

The invention provides a multi-modal RAG knowledge question answering method and device applied to the vertical field, relates to the technical field of artificial intelligence, supports multi-modal input of to-be-answered texts including texts, pictures, audios and videos, and makes full use of semantic information of multi-modal data. And utilizing a plurality of preset retrieval models to retrieve the knowledge fragments corresponding to the to-be-answered text to obtain a plurality of retrieval results. Various retrieval results are fused through an RRF algorithm, and scores are calculated by using a target reordering model for reordering, so that the recall rate is improved. And based on the retrieval content and the preset knowledge frequency information, fine tuning sorting is further carried out, so that the method is more accurate and meets user expectation. According to the method, the natural language understanding ability of a large language model, a target reward model and related information recalled by the RAG are combined to generate optimal solution data, and meanwhile, a target scoring model and domain expert knowledge are utilized to provide more professional high-quality solution data meeting actual requirements.
Owner:LINKAGE TECHNOLOGY (NANJING) CO LTD

Indoor design method and system based on artificial intelligence

ActiveCN119026212AMeet personalized interior design needsAvoid the problem of insufficient personalizationGeometric CADInternal combustion piston enginesSensing dataGenerative adversarial network
The invention provides an indoor design method and system based on artificial intelligence, and relates to the field of data analysis processing, and the method comprises the steps: obtaining indoor space sensing data and user historical behavior data; constructing an indoor space three-dimensional model according to the indoor space sensing data, and extracting indoor space features from the indoor space three-dimensional model; determining target interior design elements preferred by the user according to the historical behavior data of the user; inputting the indoor space features and the target interior design elements into a generative adversarial network to determine a plurality of corresponding preliminary interior design schemes; and inputting the plurality of preliminary interior design schemes into a multi-objective optimization model to determine a target interior design scheme, the multi-objective optimization model adopting an improved particle swarm optimization algorithm. Therefore, by combining the user preference and the indoor space characteristics, the generated design scheme can meet the high-standard requirements of the user in the aspects of demand individuation and practical functionality.
Owner:Shenzhen Habitat Technology Co., Ltd.

RAG-oriented multi-branch hybrid balance retrieval optimization method and system

The invention discloses an RAG-oriented multi-branch hybrid balance retrieval optimization method and system, and relates to the technical field of RAG retrieval, and the method comprises the steps: extracting a semantic vector of a user question, carrying out the similarity matching of the semantic vector of the question and the semantic vector of each text in a knowledge base, and generating a first candidate text set; extracting keywords in the user question, performing secondary filtering to obtain a keyword list, and performing keyword matching on the keyword list and each text in the knowledge base to generate a second candidate text set; combining the first candidate text set and the second candidate text set, performing duplicate removal, and inputting the first candidate text set and the second candidate text set into a Reranker model to generate a third candidate text set; judging whether the third candidate text set is empty or not, if not, outputting the third candidate text set as a retrieval result, otherwise, performing secondary retrieval, and re-determining and outputting the retrieval result; and sending the user question and the retrieval result to the large language model. According to the method and the device, the accuracy of RAG retrieval can be improved, and the user experience is improved.
Owner:山东省大数据中心

Air purification method and system fused with AI intelligent big data analysis

The invention relates to an air purification method and system fused with AI intelligent big data analysis, and solves the problem that the purification effect is greatly reduced due to the fact that an air purifier cannot adjust operation parameters in time according to upcoming air quality changes in the operation process. The method comprises the following steps: inputting real-time gas related data and gas related data in a historical past time range into a pre-trained air quality prediction model, and outputting an air quality prediction result in a future preset time range; and the mode set by the user about the air purifier and the air quality prediction result in the future preset time range are input into the air purifier operation parameter optimization model, the optimized air purifier operation parameters are output, and the air purifier is indicated to perform corresponding operation according to the optimized air purifier operation parameters. The intelligent air purifier has the following effects that precise and intelligent air purification is achieved, the indoor air quality is effectively improved, and various requirements are met.
Owner:北京三五二环保科技有限公司

Data synthesis method and training method for monocular three-dimensional reconstruction based on synthesized data

The invention relates to the technical field of three-dimensional reconstruction, and provides a data synthesis method and a training method for monocular three-dimensional reconstruction based on synthesized data, and the method comprises the steps: 1, rendering a picture containing a specific object into a plurality of pictures I corresponding to the object at different angles; step 2, obtaining a mask m of the object based on the picture I, and 3, guiding the picture generation model to generate a new picture which conforms to the text description and retains the structural information of the specific object by utilizing the picture I by utilizing the mask m and the additionally input text description. According to the method, a large number of 3D models and picture pairs are generated through different viewpoints, styles and background conditions, a rich training data set is provided, and the generalization ability of the monocular three-dimensional reconstruction model can be remarkably improved through the diversity. By randomly applying disturbance within a certain angle range of a plurality of viewpoints (R, T), the adaptive capacity of the model to different camera poses is improved, and the robustness of monocular three-dimensional reconstruction is enhanced.
Owner:MOXIN (HUZHOU) TECH CO LTD

Control method for switching process of EPS (electric power steering) power assisting mode and aligning mode

The invention belongs to the technical field of automobile steering control, and particularly relates to an EPS power-assisted mode and return mode switching process control method which is composed of a steering mode switching judgment module, a switching time determination module, a torque curve processing module and a switching compensation torque determination module. By designing the switching time length determination module, the return time is determined according to the steering angle of the steering wheel and the vehicle speed at the current moment, the expectation of a driver is better met, the expected value of the switching compensation torque is obtained through torque processing of the Bizier curve according to calculation of different switching conditions, and therefore the accuracy of the switching compensation torque is improved. The switching compensation torque at any switching moment is obtained by tracking the switching torque curve graph, so that the steering mode switching process of the vehicle under different working conditions is smoother, and the steering comfort of a driver is greatly improved; the problem of pause caused by sudden change of hand torque felt by a driver when the power-assisted mode and the aligning mode of the current electric power-assisted steering automobile are switched is solved, and the driving experience is improved.
Owner:JILIN UNIVERSITY

Aluminum alloy material, battery shell and power battery

The invention discloses an aluminum alloy material, a battery shell and a power battery, the aluminum alloy material at least comprises Al, Mn, Si, Cu, Fe and Ce, the mass percentage content of Ce is 0.3-1 wt%, the battery shell is prepared from the aluminum alloy material, and the power battery comprises the battery shell. The Ce with the specific content is creatively used as the raw material to be added into the aluminum alloy for modification treatment, aluminum grains can be reduced by half to one third, the grains can be refined, dislocation movement is effectively hindered, and then the strength and hardness of the aluminum alloy material are improved.
Owner:CHANGZHOU JINPIN PRECISION TECHNOLOGY CO LTD +1

High-concurrency deployment method, system and equipment for large model

PendingCN118981376AImprove processing throughputmeet expectationsResource allocationBiological modelsVideo memoryPage table
The invention provides a high-concurrency deployment method, system and equipment for a large model, and the method comprises the steps: receiving a user request, carrying out the analysis of the user request, and selecting a most suitable model node according to the working state and resource condition of a current model when the semaphore of the analyzed user request accords with a current limiting strategy; after the user request is forwarded to the most appropriate model node, model reasoning is carried out, and the number of video memories of the maximum key value cache is calculated in the model reasoning process; determining the size of a page frame, creating a page table by taking a request as a unit, managing key value cache in a form of a second-level page table, and determining a final page addressing range; and after reasoning is completed, a generated result is fed back to the user in real time. Based on the method, the invention further provides a high-concurrency deployment system and device for the large model. According to the method, the deployment process can be remarkably simplified, the threshold of technology implementation is reduced, and the processing throughput of a large model in an inference task can be greatly improved.
Owner:SHANDONG ZHIYANG ELECTRIC

Shenqi compound oral liquid and preparation method thereof

The invention relates to a ginseng-astragalus compound oral liquid and a preparation method thereof, belongs to the technical field of food beverages, and aims to solve the problems that some plant beverages in the market are added with a large amount of additives, the raw materials are high in cost and are not easy to become public consumer goods. The preparation method comprises the following steps: 1, carrying out hot extraction according to a solid-liquid ratio of 1: 10, cooling, and filtering to obtain a raw material extract, 3, concentrating the extract obtained in the step 2, and 4, adding fructo-oligosaccharide for blending. According to the invention, medicinal and edible raw materials are adopted, so that the beverage has the effects of tonifying middle-Jiao and Qi, nourishing blood and promoting salivation, and invigorating spleen and tonifying lung, and the raw materials are simple and easy to obtain, which means that the raw materials are relatively simple to obtain and low in cost, so that the beverage is easier to popularize and promote, more importantly, diversified health benefits are provided for consumers, and the beverage is an ideal choice for promoting healthy life; therefore, the plant beverage can become a public consumer product, and the application prospect of the plant beverage in the market is improved.
Owner:GANNAN ZHONGZANG PHARM CO LTD

User preference-based text video diffusion model training method and device

InactiveCN119672696Ameet expectationsperformance improvement helpInternal combustion piston enginesCharacter and pattern recognitionData setData mining
The invention discloses a user preference-based text video diffusion model training method and device, and the method comprises the steps: obtaining the text data of a user, inputting the text data into a text refining model for processing, and obtaining refined text data; inputting the refined text data into an initial text video diffusion model to obtain a plurality of initial videos; inputting the plurality of initial videos into a pre-trained user preference alignment model to obtain an optimal video aligned with the user preference; and training the text video diffusion model based on the optimal video to obtain a target text video diffusion model. According to the method, a video model is generated by aligning texts by using a help and security preference data set, and a reward model and an auditing model are trained by using data. And screening an optimal safety generation result as a supervision signal through the two models to train and optimize the text video diffusion model. And the safety can meet the expectation of human beings while the help of the text video diffusion model is improved.
Owner:PEKING UNIV

Intelligent image processing method and system based on AI

The invention relates to the technical field of AI, and discloses an intelligent image processing method and system based on AI.The method comprises the following steps that a user provides input data, and the input data comprise an original picture, text description and multi-modal information; using a pre-trained target detection algorithm to process object recognition in all original pictures in batches; classifying the objects identified in each picture, and marking the categories of the objects; extracting a group of standardized features of each class of objects; and for each object in each picture, according to the standardized features of the object and the keyword matching degree in the user text description. According to the method, the intention and preference of the user are accurately understood by introducing the multi-modal information and performing feature extraction on the original picture provided by the user, and important objects can be preferentially considered in the generation process based on a comprehensive scoring system of feature similarity, keyword matching degree and multi-modal influence; and the generated result better accords with user expectation.
Owner:ANHUI DUODIAN MEIHAO INTELLIGENT TECH CO LTD

Inland river shipping cargo loading and stowage method

ActiveCN119005832Ameet expectationsGuaranteed load rateInstrumentsBulk cargoOceanography
The invention belongs to the technical field of inland river shipping management, and particularly relates to an inland river shipping cargo loading and stowage method. Comprising the following steps: classifying to-be-transported objects according to mass and volume characteristics of the to-be-transported objects, determining an optimal loading scheme by taking a loading rate and a yield rate of an inland river bulk cargo ship as an optimal loading optimization model, and determining a cargo stowage scheme of each bay position by each port loading plan to complete stowage. According to the method, the characteristics of inland river shipping are fully considered, the characteristic that the inland river shipping is highly managed by a ship owner is considered, and a scheme for further optimizing assembly and stowage based on the highest loading rate and profit rate of a ship and the characteristics of assembly efficiency of each port is provided; according to the scheme, the optimal loading and stowage scheme of the ship on the planned route can be achieved within a shorter time, the ship loading rate is guaranteed to the maximum extent while the expectation of a ship owner is better met, and the cargo transportation efficiency in the route is improved.
Owner:THREE GORGES ELECTRIC POWER JINZHOU ENERGY (HUBEI) CO LTD

Confirmation method for accelerating on-chain transaction, automatic selling method and system

The invention discloses a confirmation method for accelerating chain transaction, and an automatic selling method and system. The confirmation method can comprise the following steps: monitoring a transfer event; sending a query request to a plurality of nodes so as to enable the plurality of nodes to perform query request feedback, and further obtaining confirmation information corresponding to the plurality of nodes; and judging whether the number of the confirmation information exceeds a preset threshold, and if yes, determining that the transaction is valid. According to the automatic selling method and system, the confirmation method for accelerating the chain transaction is applied, and compared with a transaction confirmation method which needs to wait for a majority of nodes to pass through a plurality of blocks in the prior art, the transaction confirmation time is greatly shortened, and therefore the real-time performance of automatic selling is improved.
Owner:YUANBI INNOVATION TECHNOLOGY CO LTD

Recommendation system feature fusion method and device based on reinforcement learning agent

The invention relates to the technical field related to intelligent recommendation, in particular to a recommendation system feature fusion method and device based on a reinforcement learning agent. The method comprises the following steps: acquiring a user log; respectively constructing a click rate estimation model and a reading duration estimation model based on the user log; wherein the click rate estimation model is used for estimating the click rate of an article or a video, and the reading duration estimation model is used for estimating the reading duration of the article or the video; constructing a fusion model of an output result of the click rate prediction model and an output result of the reading duration prediction model; wherein the fusion model is further used for adjusting the internal parameters based on the latest operation of the user, so that the fusion value output by the fusion model better conforms to the expectation of the user; wherein the click rate is positively correlated with the fusion value, and the reading duration is positively correlated with the fusion value; and performing recommendation based on the click rate estimation model, the reading duration estimation model and the fusion model.
Owner:BEIJING WANGZHI TIANYUAN BIG DATA TECH CO LTD +1

Exhaust system for internal combustion engines

The present invention provides an exhaust system (6) for an internal combustion engine (4) having at least one exhaust pipe (7) originating from the internal combustion engine (4) and having an end portion (9) terminating at an outlet orifice (10) through which exhaust gases are released into the atmosphere. The end portion (9) of the exhaust pipe (7) has at least one movable wall (16) movable to different positions to change the width of the outlet orifice (10). The system includes: a motor-driven actuator (17) configured to move the movable wall (16) and capable of being electronically controlled; and a control unit (17) configured to change the position of the movable wall (16) by controlling the actuator (17) according to the rotational speed and engine load of the internal combustion engine (4).
Owner:FERRARI SPA

Information collecting and editing method and device based on reinforcement learning

The invention relates to the technical field related to intelligent recommendation, in particular to an information collecting and editing method and device based on reinforcement learning. The method comprises the following steps: acquiring an intelligence collecting and editing log; based on the information collecting and editing logs, respectively constructing a selection button click rate estimation model and a push button click rate estimation model; wherein the newspaper selection button click rate estimation model is used for estimating the newspaper selection button click rate of the intelligence, and the push button click rate estimation model is used for estimating the push button click rate of the intelligence; constructing a fusion model of an output result of the newspaper selection button click rate prediction model and an output result of the push button click rate prediction model; wherein the fusion model is a deep learning model, and is also used for adjusting internal parameters based on the latest operation of the information acquisition and editing personnel, so that a fusion value output by the fusion model better conforms to the expectation of the information acquisition and editing personnel; and performing candidate intelligence recommendation based on the selection button click rate estimation model, the push button click rate estimation model and the fusion model.
Owner:BEIJING WANGZHI TIANYUAN BIG DATA TECH CO LTD +1

Scene image intelligent acquisition method and device, equipment and storage medium

The invention relates to the technical field of images, in particular to a scene image intelligent acquisition method and device, equipment and a storage medium, and the method comprises the steps: obtaining a shooting demand sent by a user; acquiring a video frame from a shot video based on the shooting demand; judging whether the video frame is matched with the shooting demand or not based on a scene feature recognition model; and if yes, acquiring a scene image corresponding to the video frame for a user to select and view. According to the invention, the efficiency of obtaining the expected scene image can be improved, and the obtained scene image can meet the individual demands of people.
Owner:DDPAI TECH CO LTD

A game virtual character behavior decision-making method based on deep reinforcement learning

PendingCN122251863AAccurate processingAccurate behavioral decisionsBiological modelsVideo gamesSimulationDecision taking
The application discloses a game virtual role behavior decision-making method based on deep reinforcement learning, comprising the following steps: S1, obtaining the current environment state information of a game virtual role, and obtaining standardized environment state features after preprocessing the environment state information; S2, performing hierarchical reinforcement decision-making based on an intention-risk double-layer decoupling architecture, and generating a candidate behavior action set matching the current environment state; and S3, performing robustness screening on the candidate behavior action set through opponent counterfactual simulation, and the application relates to the technical field of game artificial intelligence. The game virtual role behavior decision-making method based on deep reinforcement learning performs robustness screening on the candidate behavior action through opponent counterfactual simulation, ensures that the selected behavior not only conforms to the current environment state, but also exhibits good stability and benefits under various possible opponent strategies, and greatly improves the ability of the game virtual role to cope with complex situations.
Owner:SHANGHAI YOUWAN NETWORK TECH CO LTD

Multi-modal input music score generation method and device based on music and / or graph

The invention provides a multi-modal input music score generation method and device based on a music and / or graph, and the method comprises the steps: selecting node information of music data, the node information can be music segments, phrases, sections or notes, obtaining music attribute features corresponding to the node information, and generating a music score according to the music attribute features; the music attribute features can be melody, accompaniment, harmony or rhythm, the target music is generated based on the node information and the music attribute features matched with the node information, the problems of low quality and poor controllability of music generation in the prior art are solved, the method allows a user to accurately control the music generation process according to specific requirements, and the user experience is improved. Comprising the selection of styles, emotions, rhythms, timbres and the like, so that the generated music better conforms to the expectation of the user, and the user can obtain music works conforming to personal taste and emotional expression.
Owner:BEIJING INSTITUTE FOR GENERAL ARTIFICIAL INTELLIGENCE

Smart visual prosthesis system and coding method

PendingCN122075932Anatural contrastrich in detailsPhotometryInternal electrodesIlluminanceVisual prosthesis
This invention discloses an intelligent visual prosthesis system and encoding method. The system includes an external sensing processing unit and an internal implanted stimulation unit. The external sensing processing unit includes at least an environment sensing unit and a first data processing unit, and the internal implanted stimulation unit includes at least a stimulation signal generation unit and an electrode array. The environment sensing unit is used to acquire a target image. The first data processing unit is used to match the current scene to a preset lighting scene mode according to the image features or ambient illuminance of the target image, independently configure anchor points for the corresponding image grayscale value boundaries for the matched lighting scene mode, and generate a grayscale-charge density mapping relationship corresponding to the lighting scene mode. Then, the grayscale value of each pixel of the target image is mapped to the corresponding target charge density, and a stimulation command is generated and output to the stimulation signal generation unit. The stimulation signal generation unit is used to generate an electrical stimulation pulse sequence according to the stimulation command and a preset reference pulse width and output it to the electrode array.
Owner:HANGZHOU NANOCHAP ELECTRONICS CO LTD

Intelligence collection and editing method and device based on reinforcement learning

The present application relates to the technical field related to intelligent recommendation, and specifically to a method and device for intelligence collection and editing based on reinforcement learning. The method includes: obtaining intelligence collection and editing logs; constructing a selection button click rate estimation model and a push button click rate estimation model based on the intelligence collection and editing logs; wherein the selection button click rate estimation model is used to estimate the selection button click rate of intelligence, and the push button click rate estimation model is used to estimate the push button click rate of intelligence; constructing a fusion model of the output results of the selection button click rate estimation model and the output results of the push button click rate estimation model; wherein the fusion model is a deep learning model, and is also used to adjust internal parameters based on the latest operations of intelligence collection and editing personnel, so that the fusion value output by the fusion model is more in line with the expectations of intelligence collection and editing personnel; candidate intelligence recommendations are made based on the selection button click rate estimation model, the push button click rate estimation model and the fusion model.
Owner:BEIJING WANGZHI TIANYUAN BIG DATA TECH CO LTD +1

Prescription and method for treating female climacteric syndrome based on cooked pseudo-ginseng and caulis spatholobi

The invention provides a formula and method for treating female climacteric syndrome based on cooked pseudo-ginseng and suberect spatholobus stem, and relates to the technical field of medicine production.The formula comprises cooked pseudo-ginseng, suberect spatholobus stem and auxiliary medicinal materials, the weight ratio of the cooked pseudo-ginseng to the suberect spatholobus stem is 2: 3, and the auxiliary medicinal materials are composed of angelica sinensis, radix paeoniae alba and radix rehmanniae preparata. According to the traditional Chinese medicine composition, the cooked pseudo-ginseng, the caulis spatholobi and the auxiliary medicinal materials are strictly selected, it is guaranteed that the raw materials are good in quality and free of mildewing and worm damage, the influence of pollution components on the medicine safety is reduced from the source, and the health of patients is guaranteed; meanwhile, the weight ratio of cooked pseudo-ginseng to suberect spatholobus stem is accurately set to be 2: 3, the total weight ratio of auxiliary medicinal materials (angelica sinensis, radix paeoniae alba and radix rehmanniae preparata) to the cooked pseudo-ginseng and suberect spatholobus stem is set to be 1: 4, and the medicinal materials are subjected to fine pretreatment including specific cleaning, airing, drying and smashing steps, so that the medicinal materials are clear and stable in component.
Owner:YUNNAN WENSHAN KUNQI PHARMA CORP LTD

Cosmetic brush design management system based on personalized customization

ActiveCN119090596AImprove personalized customization experienceImprove satisfactionCommerceManufacturing computing systemsProcess engineeringData science
The invention discloses a cosmetic brush design management system based on personalized customization, relates to the technical field of product orders, and further realizes accurate capture and satisfaction of personalized demands of users through organic cooperation of five modules. Firstly, a user type locking module obtains related cognitive data according to self-evaluation of users and accurately divides the users into different types, so that the subsequent customization process is more accurate and effective. The sample brush data monitoring module provides a rich reference data basis for personalized design by collecting and recording characteristic data of multiple groups of sample brushes and user feedback. The brush hair testing module is used for deeply analyzing and testing the powder grabbing capacity and the releasing capacity of a makeup green hand, so that a powder grabbing capacity coefficient and a releasing capacity coefficient are accurately calculated, and effective screening is carried out according to a set screening threshold value. And finally, the personalized screening module compares and analyzes the two types of cosmetic brush products and screens out cosmetic brush products which relatively meet the requirements of the user.
Owner:SHENZHEN MEIYIYA COSMETICS CO LTD

Brewing process of five-grain American wine

The invention relates to the technical field of wine brewing, in particular to a brewing process of five-grain American wine. The main raw materials comprise, by weight, 46 parts of sorghum, 8 parts of corn, 16 parts of rice, 14 parts of glutinous rice and 16 parts of wheat; comprising the following brewing processes: soaking grains, primarily steaming, braising the grains, re-steaming, spreading for cooling, adding yeast, saccharifying in a box, fermenting in a cellar, distilling to obtain wine, soaking the traditional Chinese medicinal materials for the first time, distilling for the first time, soaking the traditional Chinese medicinal materials for the second time, distilling for the second time, soaking the traditional Chinese medicinal materials for the third time, distilling for the third time, storing, oxidizing and bottling. The invention provides a brewing process of five-grain American wine, and aims to develop high-quality medicinal liquor which not only has the flavor of the traditional five-grain white wine but also is rich in health-care components of traditional Chinese medicinal materials by organically combining the traditional five-grain white wine with the traditional Chinese medicinal materials through a unique raw material ratio and a brewing process.
Owner:HUIZHOU GUANGRENTANG LIQUOR CO LTD

A dynamic control system and method for cigar flavor

This invention discloses a dynamic control system and method for cigar flavor, relating to the field of cigar production technology. It includes: a flavor analysis module for analyzing volatile components in cigar tobacco leaves, performing sensory evaluation of the overall cigar flavor, and establishing a cigar flavor fingerprint; a demand acquisition module for collecting current consumer demand, gathering consumer personal information and cigar consumption habit data; and a consumer habit database for storing consumer personal information and cigar consumption habit data, and establishing a correlation model between consumer habits and cigar flavor characteristics through data mining and analysis techniques. The system of this invention can generate personalized cigar recipes based on consumers' personal information, consumption habits, and specific preferences, meeting consumers' personalized flavor needs and enhancing the consumer experience. Through a mathematical model of entry depth and flavor perception, combined with individual consumer habits, it refines recipe adjustment strategies, enhancing consumer participation and satisfaction.
Owner:SHENZHEN TIANJI ELECTRONIC TECH CO LTD

Mochi premixed flour, mochi and preparation method

The invention provides mochi premixed flour, mochi and a preparation method, and belongs to the technical field of food processing. The mochi premixed flour is prepared from the following components in parts by mass: 90 to 110 parts of low-gluten wheat flour, 60 to 80 parts of pre-gelatinized cassava starch, 25 to 35 parts of white granulated sugar, 1.5 to 2.5 parts of edible salt and 5 to 15 parts of whole milk powder. According to the mochi premixed flour, the low-gluten wheat flour and the pregelatinized cassava starch are used as main raw materials, protein in the wheat flour can form a gluten network, and mochi is endowed with certain toughness; the pre-gelatinized cassava starch can be gelatinized by means of high amylopectin content and high water absorption rate without heating cold water and stirring, a large amount of steam can be generated after baking to maintain the hollowness of the product, and meanwhile, the mochi has the unique taste of being soft, glutinous and elastic. A product prepared from the mochi premixed flour can keep the soft, glutinous and chewy taste of a baked mochi bag; meanwhile, after baking, the product has a relatively large volume expansion degree and is in a hollow state.
Owner:GUANGDONG PROVINCE YUEDONG TECHNICIAN COLLEGE

Image filling method and device, electronic device and medium

InactiveCN112967356BAccurate understanding of filling intentmeet expectationsImage enhancementImage analysisImage extractionComputer graphics (images)
The present disclosure provides an image filling method, device, electronic device, computer-readable storage medium and computer program product, which relate to the field of artificial intelligence, especially to the field of computer vision and deep learning technology. The implementation scheme is: obtaining an image to be filled, a first mask image and a second mask image corresponding to the image to be filled, the image to be filled includes a to-be-filled area and a background area, the first mask image indicates the relative positional relationship between the to-be-filled area and the background area, and the second mask image indicates the relative positional relationship between the user clue area and its external area; extracting background features corresponding to the background area based on the image to be filled and the first mask image; extracting clue features corresponding to the user clue area based on the image to be filled and the second mask image; calculating the similarity score between the background features and the clue features; inputting the image to be filled, the first mask image and the similarity score into a trained first neural network to obtain a filled image.
Owner:BEIJING BAIDU NETCOM SCI & TECH CO LTD

Generation management system of 3D beauty model effect library based on multi-dimensional feature data

The invention relates to the technical field of 3D beauty model effect library generation management, and particularly discloses a 3D beauty model effect library generation management system based on multi-dimensional feature data. The system comprises a 3D beauty model design module, a beauty satisfaction analysis module, a beauty improvement analysis module, a beauty stability analysis module, a beauty effect evaluation module and a beauty model effect library updating module. According to the method, the beauty effect of each beauty user is comprehensively analyzed from three aspects of beauty satisfaction, beauty improvement and beauty stability, and the 3D beauty model effect library is managed in combination with the beauty effect and access popularity of each stored 3D beauty model corresponding to each beauty demand, so that multi-angle analysis of the beauty effect is realized; and meanwhile, the 3D beauty model effect library is managed, so that repeated creation or storage of models with similar effects can be avoided, and waste of time and resources is avoided.
Owner:SHANGHAI MAISE MEDICAL TECH CO LTD

Operation data analysis method and system based on electronic commerce big data

The invention discloses an operation data analysis method and system based on e-commerce big data, belongs to the technical field of e-commerce data analysis, and aims to solve the problems that large-scale data and rapidly changing real-time data are difficult to efficiently process and hidden associated information behind the data cannot be fully mined. Comprising a data acquisition step, a data preprocessing step, a data analysis step, a result interpretation and decision support step and a data visualization step. Through a real-time data acquisition and processing technology, an enterprise can acquire latest operation data almost without delay, so that market changes and user behaviors are quickly responded, manual intervention is reduced through automatic data acquisition, preprocessing and analysis processes, the risk of human errors is reduced, meanwhile, the data processing speed is increased, and the user experience is improved. Therefore, enterprises can obtain valuable insights more quickly, and the operation efficiency and the response speed are effectively improved.
Owner:南昌理工学院