The invention discloses an intelligent traffic light control system and control method. The system comprises a cloud server side, a traffic control module and a traffic signal lamp module, the cloud server side is in wireless communication connection with the traffic control module, and the traffic control module is installed near the traffic signal lamp module at a traffic intersection and is electrically connected with the traffic signal lamp module in a wired mode; the traffic control module receives the instruction of the cloud server side, collects traffic condition data and traffic flowdata and sends the data to the cloud server side. A traffic condition type judgment standard is generated through analysis and processing of the cloud server side, a network model and a traffic lightcontrol mode are generated and returned to the traffic control module, then traffic light control signals are generated through processing of the traffic control module, and then the traffic light module is controlled to work. According to the invention, a neural network and an FPGA are dynamically configured, and a low-cost hardware system with limited computing power can be used for flexibly andefficiently controlling the traffic lights at a plurality of intersections in a certain region, so that the urban traffic operation efficiency is improved.

Intelligent traffic light control system and control method

The invention relates to a data updating method, a data updating device, electronic equipment and a storage medium. The method comprises the steps of obtaining a first image of a target object, and obtaining a first image feature of the first image; obtaining a second image feature from a local face database; performing similarity comparison on the first image feature and the second image featureto obtain a comparison result; under the condition that the comparison result is greater than a feature updating threshold, obtaining a difference feature of the first image feature and the second image feature, and taking the difference feature as a dynamic updating feature; and performing adaptive updating on the second image features according to the dynamic updating features to obtain updatedfeature data of the target object. By adopting the method and the device. The base library picture in the face database does not need to be frequently and manually updated, so that the recognition efficiency is improved.

Data updating method and device, electronic equipment and storage medium

The invention discloses a configurable universal convolutional neural network accelerator, and belongs to the technical field of calculation, calculation and counting. The accelerator comprises a PE array, a state controller, a function module, a weight cache region, a feature map cache region, an output cache region and a register stack; the state controller comprises a network parameter registerand a working state controller. An excellent acceleration effect can be achieved for networks of different scales by configuring a network parameter register, and a working state controller controlsswitching of the working state of the accelerator and sends a control signal to other modules. The weight cache region, the feature map cache region and the output cache region are each composed of aplurality of data sub-cache regions and used for storing weight data, feature map data and calculation results. According to the method, proper data reuse modes, array sizes and the number of sub-cache regions can be configured according to different network characteristics, and the method is good in universality, low in power consumption and high in throughput.

Configurable universal convolutional neural network accelerator

The embodiment of the invention discloses an identity recognition system, a client, a server and a method. The client comprises an image collection module used for obtaining a to-be-used image; a feature extraction module used for performing image feature extraction on the to-be-utilized image by utilizing a feature extraction model locally stored in the client to determine features of the to-be-utilized image; a determining module used for determining a speculated identity label corresponding to the image feature to be utilized by utilizing the image feature to be utilized and the establishedlabel corresponding relationship; an obtaining module used for obtaining a calibration identity label corresponding to the to-be-utilized image feature; a first sending module used for sending the to-be-utilized image features, the speculated identity label and the calibrated identity label to the server as training data; and an updating module used for replacing and updating the feature extraction model locally stored in the client according to the feature extraction model sent by the server so as to protect the privacy of the personnel to a certain extent and reduce the risk of leaking theprivacy of the personnel.

Identity recognition system, client, server and method

The invention discloses a convolutional neural network implementation method, system and device based on FPGA and row output priority. The method comprises the steps that editable resources of the FPGA are initialized, and an input cache module, an output cache module, a PE array module and a pooling module are generated; wherein the input caching module comprises an image caching module, a weightcaching module and a bias caching module; image data to be processed, weights and offsets are loaded through a communication module of the FPGA, and are stored in a memory storage of the FPGA; the convolution calculation is grouped, convolution calculation is carried out group by group, accumulation is carried out, and a calculation result of the whole convolution calculation is obtained; and anobtained calculation result is output through a communication module of the FPGA. According to the method, a line output priority grouping processing method is adopted, characteristic parameters can be adjusted according to different convolutional neural network model structures, and matching of hardware and software models is facilitated; based on the FPGA with limited resources, the calculationof the CNN multilayer convolution full process can be completely realized.

Convolutional neural network implementation method, system and device based on FPGA and row output priority

The invention discloses a traffic flow information acquisition terminal based on Zynq-7000, and belongs to the technical field of a traffic control system signal device The traffic flow information acquisition terminal based on Zynq-7000 uses a Zynq-7000 chip as a carrier to build an architecture including a video image acquisition sensor, an external memory module and an HDMI display, uses an AXI4 bus to internally interconnect a PS (Processing System) module and a PL (Programmable Logic) module, designs an IP core for accelerating convolutional neural network calculation, implements real-time data interaction between the PS module and the PL module by using an MCU to drive a communication architecture of an AXI4-VDMA IP core and an AXI4-DMA IP core, and integrates the functions of videoimage acquisition, storage, target detection, traffic flow statistics, output display and the like into a single chip, so that digital image processing and data transmission with a high-integration level, high-speed and low-latency can meet the real-time requirements for traffic flow statistics

Traffic flow information acquisition terminal based on Zynq-7000

The invention discloses an FPGA-based universal convolutional neural network accelerator, which comprises an MCU, an AXI4 bus interface, an address generator, a state controller, a feature map cache region, a convolution kernel cache region, a convolution calculator, and a segmentation result cache region The convolution accelerator is implemented in FPGA and contains N convolution calculation subunits; the feature map cache region and the convolution kernel cache region respectively contain N feature map sub-cache regions and N convolution kernel sub-cache regions; each convolution calculation subunit is correspondingly configured with a feature map sub-cache region and a convolution kernel sub-cache region The convolution calculator reads the data in the feature map cache region and the convolution kernel cache region and carries out convolution calculation, the calculation result of the adjacent convolution calculation subunits is subjected to multilevel accumulation; and a segmented result cache region is used for storing results of accumulation of each level output via the convolution calculator The FPGA-based universal convolutional neural network accelerator in the invention can support various convolutional neural network structures, has good versatility, requires less on-chip storage resources, and has small communication overhead

FPGA-based universal convolutional neural network accelerator

The invention discloses a video behavior identification method based on an Attention-LSTM network. The method includes transforming the input RGB image sequence through an optical flow image sequencegeneration module to obtain an optical flow image sequence; inputting the optical flow graph sequence and the original RGB graph sequence into a time domain attention frame taking module, and respectively selecting non-redundant key frames in the two graph sequences; inputting the key frame sequences of the two images into an AlexNet network feature extraction module, respectively extracting timesequence features and spatial features of the two frame images, and performing a feature weight increasing operation with strong action correlation on the feature image output by the last convolutional layer through a feature weight increasing module; and inputting the feature maps output by the two AlexNet network feature extraction modules into an LSTM network behavior identification module, respectively identifying the two pictures, and fusing the two identification results in proportion through a fusion module to obtain a final video behavior identification result. According to the invention, the function of identifying the behavior from the video can be realized, and the identification accuracy can be improved.

Video behavior identification method based on an Attention-LSTM network

The invention discloses a configurable parallel universal convolutional neural network accelerator based on BNRP, and belongs to the technical field of calculation, calculation and counting. The accelerator comprises a mode configurator, a convolution calculator, a BNRP calculator, a data communication unit and a data compression encoder/decoder. The convolution calculator comprises T pulse convolution arrays with the size of R * C, and each pulse convolution array is configured with a corresponding input and output feature map cache region and a configuration information data cache region. The BNRP calculator can execute two calculation modes and comprises R * T data input and output interfaces, R * T pooling devices, a normalization calculation module and a nonlinear activation calculation module, and all the functional modules are executed in parallel in an assembly line mode. According to the method, the execution mode of the parallel acceleration calculation module can be dynamically configured according to the characteristics of various network structures, the universality is good, the calculation complexity can be greatly reduced for the convolutional neural network with a complex network structure layer and a relatively large scale, the power consumption is low, and the throughput is high.

Configurable parallel universal convolutional neural network accelerator based on BNRP

The invention discloses an on-line learning face recognition method, belonging to the technical field of calculation and calculation, in particular to the computer vision technical field of face recognition. The method trains the facial feature extractor by using the external data set, extracting reference features corresponding to each member of the local dataset to form a reference feature space, comparing a feature vector of a sample to be tested with a reference feature to determine a reference feature that is most similar to the feature vector of the sample to be tested, when the reference feature which is most similar to the eigenvector of the sample to be tested meets the threshold requirement, taking the identity of the member whose reference feature is most similar to the eigenvector of the sample to be tested as the identity of the sample to be tested, otherwise, the message that the identification of the sample to be tested fails is returned, and the reference feature spaceis updated according to the difference between the predicted feature vector of the sample to be tested and the real feature vector corresponding to the predicted feature vector in the reference feature space, so as to adapt to the change of the facial feature over time, especially in the case of frequently changing the members.

Face recognition method based on online learning

Handle everyday research tasks with reliable, citation-backed results

Your personal Research Agent to handle research tasks with citation-backed results

Popular Tasks used by Researchers

How can I help with your research?

Meet SciSpace

Get more enhanced response by uploading the PDFs you want me to reference.

No relevant PDFs in your library

SciSpace is the AI research assistant for academics. Run systematic literature reviews on 280M+ papers, and write papers with cited sources. Trusted by 1M+ students, PhDs & researchers.

SciSpace | AI for Research

Analyze PDFs

Code & Manuscripts

Funding & Grants

Literature & Patents

Medical & Clinical Data

Systematic Review

Visualize & Present

Web & Data

Build a Google Scholar-like website for your research.

Build a website

Create charts and images for your research

Create a Chart

Write a paper for submission to a journal

Draft a manuscript

Patent Search

Design eye-catching scientific posters in minutes.

Scientific Poster Generation

Systematic Literature Review

One task is running at the moment. Your messages will be shown right after.

Drag and drop or click here to browse

Loved by <highlight>1 million+</highlight> researchers

Extract a list of specific topics and their sources from unstructured text

Topics

Compare and analyze relevant papers that matches with your search

Papers

Get insights from PDFs and bookmarked papers from your library

My library

Recent searches

Try searching for:

Catch AI-generated content in scholarly and non-scholarly content

{ai} Detector

Ai Writer

Get PDF Summaries, highlighted text explanations 

Chat with PDF

Effortlessly create in-text citations and bibliographies in APA and 2,500 other formats

Citation generator

Get explanations, summaries, and answers on academic papers

Ease up your research workflow with {scispace}'s cohort of exciting AI tools

Elevate your academic writing skills and convey your ideas the way you want

Paraphraser

Explore our range of reading and writing tools

Your file is being prepared and should be ready in a few minutes. If it's a large file, it might take a bit longer. You can close this window, and we'll email you the file when it's done.

You have reached a maximum limit of <strong>{limit}</strong> columns in the table. Remove at least <strong>1</strong> column to add or create another one.

Pang Wei

Author Tools

Chat about Author