502 results about "DataOperation" patented technology

A smart watch is small and exquisite, and has very high data operation and processing capacity; the disadvantages and the advantages of the smart watch are obvious; compared with a smart mobile phone, a display screen is small in size, so that the smart watch has great limitation in information display and user input, and is not high in practicability; and compared with the smart mobile phone or other electronic equipment, the smart watch clings to skin of a user, so that the smart watch has great advantages in detecting and utilizing life biological information of the user by adding a biological inductor. According to a scheme, a gesture recognition technology based on arm electromyography is added to the smart watch, software fully uses the signal and data processing capacity of the watch, and limited hardware circuits are increased, so that the smart watch can be controlled conveniently by a gesture, the smart watch can be changed into a universal gesture recognition controller by mounting a gesture application program, the practicability and a value of the smart watch are improved and increased greatly, and seamless integration of two technologies is realized.

The invention discloses a target detection method and system based on lightweight deformable convolution. The method comprises the following steps: a deep separable feature extraction network is created as a source network; a target network is created by using an algorithm structure which is the same as that of the source network but not performing deformable convolution layer replacement; distance loss approximation I sperformed on the output of the source network and the target network by simulating features extracted by multiple network layers of the source network; before the classification layer, the last layer of feature extraction enables the source network to approach the output of the target network layer; a new feature extraction network model is acquired after joint training through a mutual learning framework of the source network and the target network; and the new feature extraction network model is taken as a feature extractor to perform feature extraction of the image data to complete target detection. Through mutual learning of the target network and the source network, the identification precision is improved, the data operation amount is reduced, and the burden on hardware equipment is reduced.

There are provided a low power SHA-1 hash algorithm apparatus having a low power structure and optimized to a trusted platform module (TPM) applied to a mobile trusted computing environment and a low power keyed-hash message authentication code (HMAC) encryption apparatus using the low power SHA-1 hash algorithm apparatus, the HMAC encryption apparatus including: a key padder padding key data for HMAC algorithm; an XOR operator XOR operating the padded key data and a padding constant; a data connector connecting a text to be encrypted, to data obtained by the XOR operating; a data padder padding the connected data; an SHA-1 hash algorithm part performing an SHA-1 hash algorithm on the padded data; a data selector selecting and applying one of a result of the SHA-1 hash algorithm and the text to be encrypted, to the data connector; and a controller controlling operations of the key padder, data connector, and data padder, a sequence of performing a hash algorithm of the SHA-1 hash algorithm part, and storing an operation result to read data required for performing an encryption operation and store data with memory.

The invention discloses a method, a device and an application for computer numerical selection. The method comprises the following steps: confirming more than two persons capable of ensure fairness and not capable of privately communicating with each other to extract seed numbers and input the seed numbers into a computer; operating the seed numbers by the computer according to a confirmed and public seed number encryption algorithm, thereby acquiring a digital digest and publishing the digital digest; acquiring a target data by operating according to a confirmed and public target data algorithm; and then selling lottery tickets to the public or distributing application number to purchasers; and after ending, publishing a won number of the lottery or succeeded application number. The seed number encryption algorithm and the target data algorithm are both confirmed and public, the reliability of seed number can be ensured by the digital digest and the fraud possibility in the computer numerical selection can be efficiently avoided, thereby ensuring a public, fair and just result.

The invention belongs to the technical field of road traffic, and discloses a driving analog simulation system based on multi-agent traffic flow. The system comprises a BIM simulation environment module, a multi-agent traffic flow simulation module, a driving simulation platform, a control module of the driving simulation platform, a vehicle control and risk early warning module, a vehicle collision and risk identification module and a driver heart physiological data acquisition module, wherein all the modules are respectively connected to a high-performance graphic workstation for corresponding data processing, and the driver heart physiological data acquisition module and the multi-agent traffic flow simulation module are respectively connected with a data operation server and are used for storing data acquired by the sensor in real time and real traffic flow data. According to the invention, the multi-agent traffic flow simulation module realizes accurate simulation of multi-agent complex traffic flow, also realizes analysis of the driving behavior of the vehicle and the running state of the traffic flow vehicle, and finally realizes comprehensive evaluation of the traffic safety condition of the road section and the road network.

The invention provides a deep learning-oriented data processing method for GPU parallel computing. The method comprises the following steps: firstly, inputting data to model a calculation graph: (1) constructing operation rules of vertexes and edges of a directed graph; (2) defining an execution sequence of operations in the graph by using topological sorting; and (3) updating parameters through the training model, introducing a tensor life cycle, and rewriting a calculation graph based on data operation cost to obtain an optimal operation strategy, mainly comprising the following steps: firstly, modeling the calculation graph based on cost, and redefining an operation function on a CPU; fusing the swap-out operations with the same tensor into a single swap-out operation; and finally, obtaining a traversal sequence by applying a tensor replacement strategy based on calculation and transmission cost. Therefore, the computational graph modeling method based on the formalization rule is constructed. Finally, the extensible neural network and the calculation graph are combined, the training speed of the model can be increased, and the image processing effect is effectively improved.

The invention provides a household intelligent double-platform cooperative control system based on a central processing system (CPS) and a household intelligent double-platform cooperative control method based on the CPS. The household intelligent double-platform cooperative control system based on the CPS comprises a signal input module, an information processing platform, a central control platform, an executive module and a client. The signal input module is connected with the information processing platform. The information processing platform is respectively connected with the client and the central control platform through a wireless network. The central control platform is connected with the executive module. With adoption of a double-platform, complicated data operation and control operation can be conducted independently, and then operating speed of the household intelligent double-platform cooperative control system based on the CPS is improved, meanwhile, by physically integrating information of the double-platform, the executive module (physical device) and the network, multi-mode cooperative control of manual control, automatic control, manual-automatic bonding control, indoor and remote joint control and multitask real-time processing can be achieved.

The invention relates to the technical field of big data, and discloses a heat dissipation device and a heat dissipation method for a big data operation center, the heat dissipation device for the bigdata operation center comprises a shell, vent holes are formed in the top and the bottom of one side of the shell, and filter screen plates are fixedly installed in the two vent holes; a vertical hole is formed in the top of the shell, a vertical shaft is rotationally installed in the vertical hole, first fan blades are fixedly installed at the bottom end of the vertical shaft, a motor is fixedlyinstalled on one side of the top of the shell, and an output shaft of the motor is fixedly connected with one end of a driving shaft. The device is reasonable in design, ventilates a computer unit for big data operation, drives an air guide plate to swing back and forth to form swing air, conducts comprehensive and efficient heat dissipation on the computer unit, and meanwhile drives knocking plates to intermittently knock the filter screen plates for filtering dust; and dust accumulated on the filter screen plates falls off under the action of vibration, so that blockage is prevented, and the heat dissipation effect is guaranteed.

The embodiment of the invention provides a user grouping processing method and device based on machine learning and an electronic terminal, and relates to the technical field of data processing. According to the embodiment of the invention, a plurality of desensitization user data after desensitization processing are acquired; then, a decision tree model in a Spark big data operation platform is used for carrying out classification processing on the desensitization user data, and a classification processing result is obtained, so that, based on the decision path of the decision tree and the classification processing result, the class group information of each node in the decision tree model is counted layer by layer; and then, generating a user data distribution tree according to the decision path of the decision tree and the class group information of each node in each decision tree model, and storing the user data distribution tree. Therefore, according to the technical scheme provided by the embodiment of the invention, the problems of high labor cost and low user grouping accuracy of a manual user portrait grouping mode in the prior art can be solved.

The invention discloses a trusted computing method and system, and the method comprises the steps that a data user generates a data operation task in an authorization range after the validity verification of a trusted computing environment and a trusted computing function library of a cloud server is passed, and transmits the data operation task to a computing service provider; the computing service provider obtains the ciphertext of the required data and the ciphertext of the corresponding key according to the data operation task, and stores the ciphertext of the required data and the ciphertext of the corresponding key to the cloud server; a data user adopts a safe transmission mode to send a calculation private key for decrypting the ciphertext of the secret key to the cloud server, triggers the cloud server to obtain a plaintext of data required for task execution by using the calculation private key in the trusted computing environment, executes the data operation task by using the plaintext and the trusted computing function library, a corresponding execution result is encrypted and sent to the data user. According to the invention, the overall operation performance can be effectively improved.

The invention discloses a quick sectioning method based on point cloud, an intelligent terminal and a cloud platform, and the method comprises the following steps: uploading point cloud data, so as toenable the cloud platform to convert the point cloud data into an octree binary system for storage, and obtaining a point cloud file; after a viewing instruction is obtained, asynchronously loading the point cloud file from the cloud platform according to the viewing instruction, and visually rendering the point cloud according to the point cloud file; sensing the area of interest and the parameters, so that the cloud platform evaluates the number of point clouds in the region of interest and returns points; uploading a generation instruction to enable the cloud platform to calculate the number of point clouds in the area of interest to obtain target data; and uploading the downloading instruction, so that the cloud platform converts the target data into a corresponding output data type and returns the output data type. Based on a browser, a PC terminal, a mobile terminal and other clients are supported, the cloud platform stores the point cloud data and performs data operation, the hardware requirements of the clients are reduced, and the invention can be widely applied to the field of three-dimensional visualization.

The invention discloses a convolutional neural network accelerator based on a Winograd sparse algorithm. The convolutional neural network accelerator comprises; a control module which is used for moving data; a buffer module buffer which is used for temporarily storing load data, and an operation module which is used for completing operation of a Winograd sparse algorithm. In the reading stage, the control module sends an address, an input cache and a weight cache to read data in the external DRAM; in the data operation stage, the operation module reads input data, weight data and a weight index from a buffer module to complete convolution operation; and in the sending stage, when the output finishes the final accumulation operation, the data is sent to the external DRAM through the outputcache, and finally calculation is finished. The convolutional neural network accelerator has the advantages of simple structure, easiness in implementation, good acceleration effect and the like.

The invention discloses a data operation and private transaction method based on a blockchain and a trusted calculation network. The method comprises the steps that a requester provides a data operation request and only obtains a data operation result in the whole data operation process, and does not touch the original data of a provider; on the premise of authorization of a provider, the platformperforms data operation in a trusted execution environment and outputs a data operation result, and the data operation process is real and trusted; after the data operation is completed, a provider can delete authorized data in the data operation or encrypt and store the data, so that data leakage is avoided; according to the method, original data used for statistics is only visible to a provider, an obtained data operation result is only visible to a requester, data operation is executed in a trusted environment, the privacy protection requirement in the data operation process can be met, and a real and reliable data operation result is obtained.

The invention relates to a food quality monitoring device based on a neural network. The food quality monitoring device is formed by connecting a data processing and driving module with a storage and display module. The data processing and driving module is used for collecting and processing signals of a temperature receiving end and providing the capacity of driving a display and comprises the temperature receiving end, a data conversion module, a data operation module, a data processing module and a driving, scanning and displaying module. The storage and display module is formed by connecting a data storage module with the micro-display and used for displaying current temperature data and food quality coefficients in the data storage module on the display. The food quality monitoring device is characterized in that complex formula operation is achieved by adopting the neural network, a parallel processor FPGA is used, the data is transmitted to a monitoring module in a wireless transmission communication mode, and the alarm prompting function is added. According to the food quality monitoring device, not only are hardware resources saved, but also the operation and processing speed is high, the timeliness is high, and the good practical value is achieved.

This invention needs no priority encoder to connect the plural number of associative memories. The primary searching associative memory 4 of the associative memory 204 produces the intermediate data 6 obtained after the logical sum operation for the coincident storage data in the confirmed valid state, taking the research data 2 and the mask information into account, into the intermediate data determination section 41 and the internal secondary searching associative memory 5. The intermediate data arithmetic section 42 produces the valid state to the valid search signal 45 corresponding to the storage data with the least number of bits in invalid state among the intermediate data 6 supplied from the first through p-th associative memory 204. The first through p-th secondary searching associative memory 5 carries out the search operation for the storage data with the corresponding intermediate data 6 as the search data, supplying the match line 3. When the corresponding valid search signal 45 is put in the invalid state, the first through p-th logical sum operation means 46 puts the all values of corresponding match line 3 in the invalid state, produces data into the address signal producing section 11, and encodes data to the address output signal 12.

The invention belongs to the electronic weighing field and relates to a new calibration method of an electronic belt scale and supporting devices for implementing the method. According to the new calibration method of the invention, an electronic belt scale calibration algorithm is simplified, so that the electronic belt scale calibration method can be simplified, and therefore, the long-term stable and precise operation of the electronic belt scale can be realized, and the overall level of the electronic belt scale can be improved. According to the electronic belt scale, an electronic belt weighing controller with a one-key calibration function is adopted; and the weighing controller includes a main controller for ensuring high speed and high accuracy of data calculation, a weighing sensor for acquiring output signals, an instrumentation amplifier for amplifying acquired signals, a display screen which queries and modifies the internal parameters of instruments to complete calibrationcontrol, a digital quantity input and output port which can realize the programmability of different kinds of field control logic, an analog quantity input and output port which sets other analog variables, an extended installation slot for installing more functional modules, and one-key calibration system for embedded multitasking real-time operation.