495results about How to "Increase processing rate" patented technology

The invention provides a compressed convolutional neural network-oriented parallel convolution operation method and apparatus. The method comprises the steps of determining an adopted operation mode according to input control signal convolution data shift chain length selection, accumulated offset enabling and convolution calculation enabling; and by adopting two serial shift register chains, inputting convolution data, convolution parameters and channel offset, and performing 3X3 and 1X1 convolution operation at the same time for a same input convolution data stream. According to the method, a multiplier, an accumulator, a parameter register and an offset register are added only based on original serial shift register chain-based 3X3 convolution operation; the realization method is simple; the executive efficiency is high; and the convolution operation in a compressed neural network algorithm can be effectively accelerated. According to the apparatus, a plurality of characteristic graphs can be output at the same time through simple hardware unit expansion and copying; and the apparatus has the advantages of low power consumption, high function unit utilization rate and high processing speed.

The invention discloses a heavy-duty lorry driving barrier detection and tracking method based on a binocular fisheye camera, and belongs to the technical field of traffic vehicle active safety. The method mainly comprises the steps: starting two calibrated and corrected infrared fisheye cameras, disposed at two ends of the tail of a vehicle, to synchronously collect back-up environment image information when the back-up of a heavy-duty lorry is detected, outputting a back-up information after jointing, and displaying the back-up image on a display screen of a cab. A binocular fisheye visual system is used for obtaining the initial position and distance of a to-be-tracked barrier object, and then a video target tracking algorithm combined with the low-rank matrix theory based on particle filtering is used for achieving the tracking of a barrier, achieves the real-time updating of an image template in a tracking process, and detects whether there is a new barrier or not. When a vehicle is detected to turn around, a supersonic radar is started to detect that there is a barrier in a one-meter range beside the lorry, the position of the barrier is displayed and the early warning is sent. The method is mainly used for a large-size lorry.

The invention discloses a video copy detection method and a system. The method comprises the following steps: an inquired video and a detected video are converted to digital videos with the unified format; the inquired video and the detected video are carried out the video structure analysis, shots in the videos are segmented and key frames are extracted; an index is established for the key frames of the detected video, the key frames of the inquired video are utilized for carrying out the preliminary search of a video copy in the index, the detected video containing the similar content with the key frames of the inquired video is obtained, and the detected video is taken as the option video for the next matching; the similarity between the key frames of the inquired video and the sequence of the key frames of the option video is calculated, the matching of the sequence of the key frames is carried out, the similarity between the videos are confirmed, and the video with the high similarity is taken as a copy video source of the inquired video. The video copy detection method and the system can reduce the search space for the video copy detection and effectively improve the detection efficiency.

The invention discloses a method for displaying application programs and a mobile terminal. The method includes steps of determining whether second application program floating window quick startup operation instructions inputted by users are acquired or not by means of monitoring when the first application programs run; starting up the second application programs in backgrounds if the second application program floating window quick startup operation instructions inputted by the users are acquired, and outputting and displaying floating windows of the second application programs on display interfaces of the first application programs. The method and the mobile terminal have the advantages that the floating windows of the second application programs can be directly outputted on the display interfaces of the first application programs after the operation instructions are acquired, accordingly, switching and utilization of Home keys can be omitted, the use experience can be enhanced for the users, and the service lives of the Home keys can be prolonged; required data can be directly acquired from the display interfaces of the first application programs by the users and can be inputted into the second application programs, and accordingly the processing efficiency can be enhanced.

The invention a data rarefying and smooth processing method based on a vector map. The method mainly comprises the following steps: the data volume is reduced by rarefying curve characteristic points, and the protruded edge problem is solved by smoothly processing closed curve characteristic points. The curve characteristic points are compressed by setting different segment length threshold values, angle threshold values and vertical distance threshold values, and unsmoothed broken lines are smoothly processed through the smooth processing algorithm, so the display effect of closed curves having disadvantages of large curvature change and protruded edges can be effectively improved.

The invention discloses a method and a device for supervising traffic based on a token bucket. The device comprises a setting module, a memory, a control module and a traffic supervising module. The corresponding method comprises the following steps that: 1) setting different leaky bucket algorithm rules, and establishing a mapping relationship table between stream IDs and the leaky bucket algorithm rules; and 2) acquiring a message to be supervised; searching for the corresponding leaky bucket algorithm rule and a leaky bucket parameter by using a stream ID as an index; accordingly determining the number of available tokens in the current average leaky bucket and the peak leaky bucket; comparing the length of the message with the number of the available tokens in the current average leaky bucket and the peak leaky bucket respectively; coloring the message by adopting a corresponding leaky bucket algorithm rule and according to the comparison result; updating and storing the leaky bucket parameter serving as a standby parameter for processing the next stage of message with the same stream ID; and then discarding or forwarding the message according to the coloring result. The method and the device simultaneously support a single-rate marker algorithm and a double-rate marker algorithm, have excellent compatibility, break through the bottle neck of processing capacity and have high processing rate.

The invention discloses a hardware LZMA (Lempel-Ziv-Markov chain-Algorithm) compression system. The hardware LZMA compression system comprises a PCIE (peripheral component interface express) interface module, a direct entry data access module DMA (direct memory access), a data read-in cache module, an LZ77 compression encoding module, a section encoding module, a data read-out control module and a direct outgoing data access module DMA, wherein the PCIE interface module is connected and communicated with an upper computer, the direct entry data access module DMA is used for directly accessing data, the data read-in cache module is used for caching data to be compressed, the LZ77 compression encoding module is used for compressing and encoding the data to be compressed by the LZ77 algorithm and generating correlation control signals of the data read-in cache module and the section encoding module, the section encoding module is used for encoding sections and compressing the LZ77 compressed data again, the data read-out control module is used for connecting the compressed data outputted by the section encoding module into data types more suitable for external high-speed buses and caching the connected data, and the direct outgoing data access module DMA is used for directly accessing data. By the hardware LZMA compression system, processing speed of the LZMA compression algorithm on the basis of software is effectively increased, and a CPU is disengaged from compression of massive data.

Disclosed is a method for enabling user equipment (UE) to steer traffic in a wireless communications system. The method comprises the steps of: transmitting UE traffic steering preference information to a first access network; receiving traffic steering information from the first access network, the traffic steering information being set based on the UE traffic steering preference information; evaluating the traffic steering based on at least one of a first and a second traffic steering rule; and performing the traffic steering between the first access network and a second access network based on the result of the evaluation of the traffic steering. The first traffic steering rule is prescribed by ANDSF (Access Network Discovery and Selection Functions). The second traffic steering rule is prescribed by the first access network.

The invention discloses a coupling filler autotrophic denitrification biofilter and application and belongs to the technical field of sewage treatment. Pyrrhotite, sulfur and a carbon source are utilized to be mixed and coupled according to certain proportion; then the materials are put into a reactor, and the autotrophic denitrification biofilter can be formed after inoculating biofilm formation;sewage flows through the biofilter, and nitrogen and phosphorus in the water are removed. The biofilter disclosed by the invention has the advantages of simpleness, easiness, practicability and shortbiofilm formation time; when the biofilter is utilized to treat nitrogen and phosphorus containing sewage, hydraulic retention time is only 0.5 to 3 h, and yielding water can meet a requirement; furthermore, the biofilter has the advantages of excellent nitrogen and phosphorus removal effect, low cost and suitability for engineering application.

The invention discloses a summary information extraction method and device based on a search engine and the search engine. The method comprises: obtaining matched webpage resources based on a search character string received by the search engine; identifying the webpage types of the webpage resources; extracting corresponding summary information from the webpage resources aiming at the webpage types; and outputting the summary information. According to the summary information extraction method and device based on the search engine, the situation that a user may frequently click pages corresponding to search results to find required information can be reduced, and thereby retrieval speed is improved, interaction times of the search engine are reduced, and date processing speed is enhanced.

The invention discloses a hybrid pipelining device used for video coding, which comprises a prefetched data acquisition device, a mode selector, a reconstruction loop device and an entropy coder. The reconstruction loop device comprises an intra-frame predictor, an inter-frame predictor, a differential reconstruction device and a deblock filter; data of the same macro block is processed through the prefetched data acquisition device, the mode selector, the reconstruction loop device and the entropy coder sequentially in different timeslices; and in the same timeslice, the prefetched data acquisition device, the mode selector, the reconstruction loop device and the entropy coder process the data of four different macro blocks. The hybrid pipelining device increases the PDR, reduces the area of a memory, and serves the purposes of high definition and real-time coding.

The invention relates to a natural language processing technology, and provides a training method of a named entity recognition model. The method comprises the following steps: obtaining unlabeled corpus data from a corpus; pre-tagging the corpus data based on a category thesaurus; modifying the pre-labeling result to obtain a training text; training the named entity recognition model based on thetraining text, and repeating the above steps until the accuracy of the named entity recognition model reaches a preset threshold. Based on the training method provided by the embodiment of the invention, the recognition result of the named entity recognition model is more accurate and meets the actual requirement. In addition, the invention also provides a training system of a named entity recognition model.

The invention relates to a digital terminal supporting Braille input and a Braille input method, wherein the digital terminal comprises a key input unit including a Braille position input key group and an operation functional key group, a key identification unit, a Braille processing unit, an application layer unit, a character and speech synthesis unit, an audio frequency drive unit and an audio frequency generating unit, wherein the audio frequency driving unit is used for driving the audio generating unit to generate audio frequency. The digital terminal supporting Braille input and the Braille input method can realize barrier-free character communication between the blind and persons with vision disorder, thereby being beneficial to popularization and application.

The invention provides an optimal electrode assembly automatic selecting method of a brain-machine interface, which relates to the field of brain-machine interfaces. According to the invention, the automatic selection of the optimal electrode is realized; and the optimal electrode assembly automatic selection of an imagery motion brain-machine interface is realized by using a combined method of a common special pattern (CSP) and a support vector machine (SVM). The automatic selection of the optimal electrode assembly in the brain-machine interface has important effects on simplifying a brain-machine interface system and increasing system classification recognition rate and data transmission efficiency; and by using relevant property of an optimal support vector machine SVM linear kernel function as the optimal electrode screening index, redundant electrodes can be effectively deleted, the number of the electrodes can be obviously reduced, the useful electrodes can be kept, the system performance can be improved, and the precondition can be created to improve the popularization of a brain-machine interface technology.

The invention discloses a voice synthesis model training method and system, a voice synthesis method and system, equipment and a medium. The training method comprises the following steps: acquiring aplurality of pieces of historical text information and historical speech information thereof; obtaining a historical text vector of the historical text information; constructing an initial acoustic model based on a CNN and a bidirectional LSTM network; based on the historical text vector and the first Mel spectrum of the historical voice information, performing model training on the initial acoustic model to obtain a target acoustic model; and based on the second Mel spectrum and the historical voice information, performing model training on a preset neural network model to obtain a target vocoder model. According to the invention, the acoustic model is built based on the CNN, the bidirectional LSTM network and a linear layer, and the vocoder model is built based on a generative adversarial network GAN, thereby greatly improving the voice synthesis speed while guaranteeing the voice synthesis quality, and meeting the demands of an e-commerce platform for a large number of outgoing calls.

The invention discloses a sludge heating and drying system. The sludge heating and drying system comprises a pretreatment mechanism, a transmission mechanism and a drying mechanism, wherein the pretreatment mechanism comprises a dehydrating cavity and a dispersing cavity, a dehydrating drum and an extrusion device are arranged in the dehydrating cavity, a tilting table and a dispersing device are arranged in the dispersing cavity, and the drying mechanism comprises a drying box body and a heating device. The drying method comprises the following steps: (1) carrying out primary dehydration treatment on sludge; (2) carrying out dispersion treatment on the sludge; (3) carrying out drying treatment on the sludge; and (4) cleaning, checking and repairing. The sludge heating and drying system disclosed by the invention is simple in structure and strong in practicability, most of sewage in the sludge is removed by virtue of the dehydrating cavity, then the dispersion treatment is carried out on the sludge extruded into blocks by virtue of the dispersing cavity, working pressure of the drying box body is reduced, sludge drying effect and sludge drying quality are improved, water resource is saved, and comprehensive utilization rate is improved.

The invention discloses ETL (extract-transform-load) data processing method, device and system. The method includes: acquiring source data from a data source, and converting the source data into source data in the CSV (comma-separated value) format; establishing an ETL task according to the source data in the CSV format, and triggering an ETL tool to run the ETL task; transmitting running results of the ETL task to a target data warehouse. The data, to be processed, of the data source is uniformly converted in the CSV format and is then extracted, transformed and loaded; loading results are disposed in the target data warehouse; in case of varieties of data sources, adding a conversion method of CSV-formatted data is required only; accordingly, setting of various drive / import / export tools causes decreased processing speed of ETL data in the prior art is avoided, and processing speed and efficiency of the ETL data can be higher.

The embodiment of the application discloses order and picking task processing methods, a server, a picking terminal and an electronic price tag, wherein, the order information processing method comprises the following steps: receiving order information, wherein the order information includes at least one commodity identification; sending a picking task to a picking terminal according to the orderinformation, wherein the picking task includes commodity information corresponding to the commodity identification; sending a prompting instruction to the prompting device corresponding to the commodity identification, so that the prompting device sends a prompting signal, so that a picking person using the picking terminal can determine the commodity corresponding to the commodity identificationaccording to the prompting signal. The order and picking task processing methods, the server, the picking terminal and the electronic price tag provided by the application can improve picking efficiency of a picking person, thereby increasing processing speed of order information.

The embodiment of the invention provides an online learning analysis system and method. The system comprises an image acquisition device, an image processing device, a confusion time determination device, a course video determination device, and an information pushing device. The image acquisition device is used for collecting an image of a user. The image processing device is used for processingthe image to obtain actual classes of user expression types in the image, and screening out an image with the actual class of a confusion expression. The confusion time determination device is used for determining confusion time based on the time of the image of the confusion expression. The course video determination device is used for determining a doubt course video based on the confusion time.And the information pushing device is used for pushing the content of the doubt course video to the user. With the online learning analysis system and method, the on-line learning quality can be improved.

The invention is a water treatment reactor that integrates photocatalysis and biodegradation, and a processing method thereof. A water treatment reactor integrating photocatalysis and biodegration comprises a photocatalysis plate (3) as the baffle plate between a photocatalysis section (2) at the upper part of the reactor and a biodegradation section (10) at the lower part of the reactor; an upper water hole (1) and a lower water hole (5) are opened on the photocatalysis plate; an ultraviolet source (4) is controlled by a switch and is positioned above the photocatalysis plate (3); a biomembrane carrier (8) is positioned in the biodegradation section (10); a water outlet (6) and a valve are opened on the vessel wall at the lower edge of the photocatalysis section (2); a water circulation device is arranged on the reactor (9); the photocatalysis plate shields the microbes in the biodegradation section to protect the microbes from the harm by ultraviolet light; under the joint action of the photocatalyst and the biomembrane, the organic matter that is hard to degrade can greatly increase the rate of water treatment. The reactor is characterized in compact device structure, high efficiency and high speed.

The invention provides a digital triggering system of an oscilloscope. An ADC module is used for converting analog signals from an analog channel of the oscilloscope into digital signals. A data reconstitution module is used for receiving four 8-bit 625 Mbps data streams output by the ADC module in parallel and outputting x(n) to an up-sampler module in a buffering storage mode. The up-sampler module is used for receiving signals x(n) of the data reconstitution module, so that pixel points of a screen and waveform sampling points are kept consistent. A digital comparator compares signals y(n) output by an up-sampler with a specified triggering threshold value ranging from 0 to 255, an output level is changed, and triggering signals are output. By the adoption of the scheme, the digital triggering system is completely achieved in an FPGA; by the adoption of the digital interpolation mode, the precision of time interval measurement is improved, triggering jittering of the oscilloscope is reduced, and the processing speed and the refresh rate are greatly increased under the circumstance that digital post-processing compensation for jittering is not needed.

The invention discloses a Montgomery modular multiplication-based data processing method and belongs to the technical field of computers. The method comprises the following steps of: initializing a fifth random access memory through a central processing unit (CPU), and executing the following operations on the contents in a second random access memory by taking words as units, namely (1) calling a multiplying and adding module to multiply one word of the contents of the second random access memory and contents of a first random access memory and add the product with the contents of the fifth random access memory; (2) extracting one word from an operation result of the step (1) and multiplying the word with the contents of a constant register, and writing the low-order word of the multiplied result into a fourth register; and (3) calling the multiplying and adding module to multiply the contents of the fourth register with the contents of the third random access memory and add the product with the contents of the fifth random access memory; and finally, outputting the contents of the fifth random access memory according to the contents of the third memory.

The invention relates to a method for building and maintaining a trust relation between autonomy systems in an inter-domain routing system. The building and maintaining of a trust relation between autonomy systems are completed by the following steps: (1) initialization of parameters, namely defining and initializing the parameters needed to build the trust relation between the autonomy systems; (2) detection for the authenticity of an inter-domain routing message, namely judging whether the received inter-domain routing message conforms to a real network topology or not by the autonomy systems; (3) trust degree update based on a routing detection, namely updating the trust degrees in real time by the autonomy systems according to the detection result of the routing message from the adjacent autonomy system; (4) trust level judgment of the autonomy systems, namely dividing the autonomy systems into low, middle and high trust levels according to the current trust degrees of the autonomy systems, and directly discarding the routing update data packets transmitted by the autonomy systems according to the levels, or detecting the data packets, or directly updating a route list.

The invention relates to a data processing method and system of turn-based games, a client end and a server end. The method includes the following steps that: all combat data of current combat units and the position data and account data of non-combat units are only recorded through the server; the calculation of an operation sequence can be completed automatically according to operation instructions of the current combat units which are transmitted by a client end, combat results are obtained; and whether the other combat data such as blood, skills and states of the non-combat units are increased is decided according to the combat results. With the data processing method and system adopted, the quantity of data exchanged between the server and the client end can be decreased with normal proceeding of multi-turn combats ensured, and the data processing burden of the server can be reduced, and memory space in the server, which is occupied by the combat data of the non-combat units can be decreased, and therefore, the processing and response rate of the server can be increased, and experience of game players can be enhanced.

The invention discloses a quick processing method for stream meteorological data. A specific implementation process comprises the steps of acquiring the stream meteorological data in real time, wherein the stream meteorological data includes national station data and regional station data; sending the stream meteorological data to specified computing node ports; and monitoring the computing node ports, when it is monitored that the stream meteorological data passes through the computing node ports, storing the stream meteorological data in computing node memories and backing up the stream meteorological data in a distributed file system, then performing structured processing based on the stream meteorological data, designing a corresponding table structure, and establishing an index for a user to query. According to the method, the characteristics of the stream meteorological data of typical industries are fully considered, and the real-time acquisition and statistics analysis of the stream meteorological data can be realized, so that the stream meteorological data can be provided for the user to query.

The invention discloses a high-speed data frame conflict-free enqueueing processing method and a high-speed data frame conflict-free enqueueing processing device. The high-speed data frame conflict-free enqueueing processing device comprises a CPU module, an enqueueing scheduling module, a conflict detection module, a queue information management module, a cache management module, a receiving busmodule and a cache area module. The high-speed data frame conflict-free enqueueing processing method comprises the steps of: reading an enqueuing frame application request, performing threshold arbitration, allocating an idle cache address for data frames meeting requirements, and moving the data frames into a corresponding address of a cache region by means of the receiving bus; and detecting a head-of-queue information updating conflict situation according to conflicts, determining whether to update head-of-queue information in advance, correcting a queue length updating value according to aqueue length information updating conflict situation, and completing the enqueueing processing of the data frames. The high-speed data frame conflict-free enqueueing processing method and the high-speed data frame conflict-free enqueueing processing device support data frame parallel enqueueing and dequeueing processing, solve the problem of queue information updating conflicts, and accelerate the data frame enqueuing processing speed.

The embodiment of the invention provides a GUI rendering display method, a terminal, a server, electronic equipment and a storage medium, and relates to the technical field of software. The method comprises the steps that interactive operation information sent by a terminal is received, a rendering instruction of each frame of image and a timestamp corresponding to a screen updating period are updated according to the interactive operation information, a server and the terminal render the to-be-rendered frame of image independently according to the rendering instruction, and rendered frame images are obtained, and the server sends the rendered frame image to the terminal, the terminal receives the rendered frame image sent by the server, and the rendered frame image and the terminal rendered frame image are jointly used for displaying the system GUI. According to the embodiment of the invention, the frame image processing capabilities of the terminal and the server can be isolated, the terminal and the server do not influence each other, give full play to the respective capabilities, and perform synchronization according to the timestamp information to avoid time sequence errors, so that the image processing rate can be improved, hardware resources of the terminal and the server can be fully utilized, the overall efficiency is improved, and the user experience is improved.

The invention discloses a method for managing tasks, a task processing device and communication equipment in at least two aspects. The method comprises the steps as follows: a memorizer receives a task configured by an initiator access nucleus or a receiving access nucleus; and the memorizer memorizes all tasks in a manner of at least one queue, wherein each queue memorizes the tasks by adopting an FIFO (first in, first out) manner or a stacking manner or a random manner. With the adoption of the technical scheme, the task processing speed can be improved, the logic area is reduced, the cost is reduced, and the expansibility is better.

The invention discloses a sludge concentration and dewatering treatment method. The method comprises the following steps: (1) sludge pulverizing treatment; (2) sludge conditioning treatment; (3) sludge extrusion treatment; (4) sludge drying treatment; and (5) sludge storage. The method disclosed by the invention is simple in structure, high in practicality, convenient to use and remarkable in the sludge dewatering effect; through the multiple treatments of a pulverizing cavity, a conditioning tank and a dewatering cavity, void water, adsorbed water and capillary water in sludge are all separated out, so that the dewatering rate of the sludge is effectively improved, and the volume of the sludge is reduced, thereby facilitating subsequent recycling of the sludge; and the overall structure design is more reasonable and scientific.