WO2021057213A1 - 一种智能图像识别进行大数据采集分析***及应用方法 - Google Patents

一种智能图像识别进行大数据采集分析***及应用方法 Download PDF

Info

Publication number
WO2021057213A1
WO2021057213A1 PCT/CN2020/102654 CN2020102654W WO2021057213A1 WO 2021057213 A1 WO2021057213 A1 WO 2021057213A1 CN 2020102654 W CN2020102654 W CN 2020102654W WO 2021057213 A1 WO2021057213 A1 WO 2021057213A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
video stream
image recognition
module
storage
Prior art date
Application number
PCT/CN2020/102654
Other languages
English (en)
French (fr)
Inventor
陆云
吴承昊
Original Assignee
上海意略明数字科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海意略明数字科技股份有限公司 filed Critical 上海意略明数字科技股份有限公司
Priority to US17/294,105 priority Critical patent/US11483621B2/en
Priority to EP20867413.5A priority patent/EP3866057A4/en
Publication of WO2021057213A1 publication Critical patent/WO2021057213A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/94—Hardware or software architectures specially adapted for image or video understanding
    • G06V10/95—Hardware or software architectures specially adapted for image or video understanding structured as a network, e.g. client-server architectures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/4508—Management of client data or end-user data
    • H04N21/4532—Management of client data or end-user data involving end-user characteristics, e.g. viewer profile, preferences
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4662—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms
    • H04N21/4665—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms involving classification methods, e.g. Decision trees
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4667—Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections

Definitions

  • the invention relates to the technical field of data intelligent analysis, in particular to a system and an application method for intelligent image recognition to collect and analyze big data.
  • the data is easy to be missed, the error is large, the manual operation is slow, the cost is high, and the quality of manual digital behavior data collection and analysis is uneven and cannot be accurately controlled; compared with manual data collection and analysis,
  • the computer analyzes the data that the brand cares about and consumers care about, and performs modular analysis, which has the advantages of fast, high efficiency, precision, and not easy to miss that cannot be achieved by humans. Therefore, it is of great significance to provide an intelligent image recognition system and application method for big data collection and analysis in response to the above problems.
  • the purpose of the present invention is to provide a big data collection and analysis system and application method for intelligent image recognition, which are transmitted to an intelligent cloud server by means of digital behaviors converted into video streams for frame-by-frame disassembly and analysis, and the video streams are processed in an AI intelligent manner.
  • the image data of the data conversion is recognized and the analysis results are generated for application.
  • the server set in the cloud performs the management, storage, and processing of the video stream big data
  • the processed image data and training include but not limited
  • the brand classifier, advertising classifier, and product classifier models are compared and analyzed in a distributed manner, they can restore consumers' non-private real digital behaviors and produce more commercial value without delay, omission, or speed.
  • the problems of slowness, large error, and high cost solve the bottleneck of consumer real-time digital behavior analysis, bring business analysis closer to the reality, and bring more valuable analysis results to the brand to guide the brand to optimize the consumption path globally.
  • An intelligent image recognition for big data collection and analysis system of the present invention includes an intelligent cloud server, the intelligent cloud server includes a computing server and a storage server that are interactively connected, and the computing server is equipped with a data reading module connected in sequence
  • An image recognition system consisting of a video stream data processing module, an AI image recognition module, a data storage module, and a model tuning module, the storage server is provided with an interactively connected video stream storage database, a video stream management module, and a data center database;
  • the data reading module is interactively connected to the video stream storage database;
  • the intelligent cloud server receives through the communication network and the HTTP(S) protocol collected by the video stream collection terminal, including but not limited to pictures, live videos, ordinary videos and video streams.
  • the big data of data transmission is stored in the video stream storage database; the video stream storage database is used to store the acquired video stream data, and the data center database is used to store the device information on the user's video stream collection terminal, and will obtain The video stream matches the video stream collection terminal; the video stream management module is used to manage the upload, deletion, and read order of the video stream; the data reading module is used to read the order in the video stream storage database
  • the video stream data processing module is used to process the video stream data, filter and process it into image data that can be effectively recognized by the AI image recognition module; the AI image recognition module is used to process the video data
  • the image data processed by the module performs AI intelligent image recognition and compares it with the image recognition classifier model stored in the data storage module to obtain the recognition result data; the model tuning module is used for the recall rate and the recall rate of the AI image recognition module
  • the AI image recognition module is tuned to the recognition model through additional training of the classifier model and classification samples to make the recognition more accurate and detailed.
  • the communication network adopts including but not limited to 4G, 5G, WIFI network.
  • the video stream collection terminal adopts any terminal that can record the digital behavior of the user's operation and form video stream data, including but not limited to computers, smart phones, tablet computers, and smart TVs.
  • the data center database is provided with a device video stream storage path data table, a user configuration data table, and a device status data table; the device video stream storage path data table is used to combine the acquired user video stream data with the video stream. Establish and allocate storage paths among the partitions of the storage database; the user configuration data table is used to obtain and store configuration parameter data information of the user equipment that transmits video stream data; the device status data table is used to obtain and store the transmission video stream in real time The device status data information of the user device of the data, including running, interrupted, and terminated.
  • the data storage module includes, but is not limited to, an advertisement classifier model, a brand classifier model, and a product classifier model.
  • An application method of intelligent image recognition for big data collection and analysis system includes the following steps:
  • S01 Obtain video stream data.
  • the video stream collection terminal obtains the non-private and secure video stream data generated by the user during the use of the terminal by means not limited to screen recording;
  • S02 Transmit and store the video stream data, and stream the video through the communication network
  • the data is transmitted to the video stream storage database in the intelligent cloud server for storage, and at the same time, it matches the storage path of the user video acquisition terminal corresponding to the video stream storage database, stores and records the configuration data and device status data of the user video acquisition terminal;
  • S03 The intelligent cloud server reads data, reads the video stream data sorted in the video stream storage database to be read through the data read module and transmits it to the video stream data processing module for processing;
  • S04 the intelligent cloud server reads the data Take the video stream data read by the module for processing, including two kinds of processing: the first one is to directly disassemble the video key frame frame by frame of the acquired video stream data, and then disassemble the continuous time sequence obtained after frame by frame disassembly.
  • the second method is to extract the unique physical characteristics of the video stream from the acquired video stream data.
  • the video key frames of the first method are disassembled frame by frame to obtain the image recognition key frames and then sent to the AI image.
  • the recognition module performs image recognition; the intelligent cloud server automatically allocates processing methods according to the size of the read video data stream; S05: AI recognition is performed on the processed image data, according to pre-training including but not limited to advertising classifiers and brands The classifier and product classifier perform distributed recognition of the image; if the recognition is successful, the result of the corresponding parameter and the corresponding type is fed back to the intelligent cloud server; if the recognition fails, the data is discarded; S06: the acquired image after recognition Data classification and storage: Reclassify the feedback data to obtain complete consumer behavior and contact data and store it in the storage server; S07: tune the AI image recognition model: Recognize according to the AI image
  • the module has problems with recall and accuracy, and constantly adds samples and tests to optimize the AI image recognition model.
  • the physical characteristics in step S04 include color characteristics, video texture characteristics, and video motion characteristics of the video stream data.
  • the present invention transfers the digital behavior into a video stream and transmits it to the intelligent cloud server for frame-by-frame disassembly and analysis, and recognizes the image data converted from the video stream data in an AI intelligent manner and generates the analysis result for application.
  • the server set up in the cloud performs the management, storage, and processing of video stream big data, and the processed image data is compared with the trained model including but not limited to brand classifiers, advertisement classifiers, and product classifiers in a distributed manner.
  • the analysis restore consumers' non-private real digital behaviors and produce more commercial value. In this process, there will be no delays, omissions, slow speed, large errors, high costs, and solve the bottleneck of consumer real-time digital behavior analysis. , Make business analysis closer to reality, bring more valuable analysis results to the brand, and guide the brand to optimize the consumption path globally.
  • Figure 1 is a schematic diagram of the connection and structure of a computing server and a storage server of a system for intelligent image recognition for big data collection and analysis of the present invention
  • FIG. 2 is a schematic diagram of the hardware connection structure of the present invention.
  • Figure 3 is a structural diagram of the data center database in Figure 1;
  • Figure 4 is a structural diagram of the data storage module in Figure 1;
  • Fig. 5 is a step diagram of an application method of the intelligent image recognition system for big data collection and analysis of the present invention.
  • An intelligent image recognition system for big data collection and analysis of the present invention includes an intelligent cloud server.
  • the intelligent cloud server includes an interactively connected computing server and a storage server.
  • Image recognition system composed of data reading module, video stream data processing module, AI image recognition module, data storage module, and model tuning module.
  • the storage server is provided with an interactively connected video stream storage database, video stream management module, and data Central database; the data reading module is interactively connected with the video stream storage database;
  • the intelligent cloud server receives the big data collected by the video stream collection terminal through the communication network and HTTP(S) protocol, including but not limited to pictures, live videos, and ordinary videos transmitted by video stream data, and stores them in the video stream storage database;
  • HTTP(S) protocol including but not limited to pictures, live videos, and ordinary videos transmitted by video stream data
  • the video stream storage database is used to store the acquired video stream data
  • the data center database is used to store the device information on the user's video stream collection terminal, and match the acquired video stream with the video stream collection terminal
  • the video stream management module is used to manage the video The sequence of uploading, deleting and reading the stream
  • the data reading module is used to sequentially read the arranged video stream data in the video stream storage database;
  • the video stream data processing module is used to process the video stream data, filter and process it into an image that can be effectively recognized by the AI image recognition module Data;
  • AI image recognition module is used to perform AI intelligent image recognition on the image data processed by the video data processing module and compare it with the image recognition classifier model stored in the data storage module to obtain the recognition result data; model tuning module Aiming at the recall rate and correct rate of the AI image recognition module, the AI image recognition module is tuned by additional training of the classifier model and classification samples to make the recognition more accurate and detailed.
  • communication networks include but are not limited to 4G, 5G, and WIFI networks.
  • the video stream collection terminal adopts any terminal that can record the digital behavior of the user's operation and form video stream data, including but not limited to computers, smart phones, tablet computers, and smart TVs.
  • the data center database is provided with a device video stream storage path data table, a user configuration data table, and a device status data table;
  • the device video stream storage path data table is used to combine the obtained user video stream data with the partitions of the video stream storage database.
  • the user configuration data table is used to obtain and store the configuration parameter data information of the user equipment that transmits video stream data;
  • the device status data table is used to obtain and store the device status data information of the user equipment that transmits video stream data in real time , Including running, interrupted, and terminated.
  • the data storage module includes, but is not limited to, an advertisement classifier model, a brand classifier model, and a product classifier model.
  • an application method of a big data collection and analysis system for intelligent image recognition includes the following steps:
  • S01 Obtain video stream data, and the video stream collection terminal obtains the non-private and secure video stream data generated by the user during the use of the terminal in a way that is not limited to screen recording;
  • S02 Transmit and store the video stream data, transmit the video stream data to the video stream storage database in the intelligent cloud server through the communication network for storage, and match the user's video acquisition terminal corresponding to the storage path, storage and storage in the video stream storage database. Record user video to obtain terminal configuration data and equipment status data;
  • the intelligent cloud server reads data, reads the video stream data sorted in the video stream storage database to be read through the data read module and transmits it to the video stream data processing module for processing;
  • the intelligent cloud server processes the video stream data read by the data reading module, including two types of processing:
  • the first method is to directly disassemble the video key frames of the acquired video stream data frame by frame, then perform differential analysis and comparison on the continuous time sequence images obtained after frame by frame disassembly, and select and stitch the required images.
  • the identified key frames are sent to the AI image recognition module for image recognition;
  • the second method is to extract the unique physical characteristics of the video stream from the acquired video stream data, perform differential analysis and comparison with the detection model set in the storage server, and select useful video stream data.
  • the video key frame of the method is disassembled frame by frame to obtain the image recognition key frame and then sent to the AI image recognition module for image recognition; the intelligent cloud server automatically allocates the processing method according to the size of the read video data stream;
  • S05 Perform AI recognition on the processed image data, and perform distributed recognition of the image according to pre-trained, including but not limited to advertising classifiers, brand classifiers, and product classifiers; if the recognition is successful, it will feed back the corresponding parameters to the intelligent cloud server The result and corresponding type; if the recognition fails, the data will be discarded;
  • S06 Classify and store the acquired image data after recognition: reclassify the feedback data to obtain complete consumer behavior and contact data and store it in the storage server;
  • the physical characteristics in step S04 include color characteristics, video texture characteristics, and video motion characteristics of the video stream data.
  • the present invention transfers the digital behavior into a video stream and transmits it to the intelligent cloud server for frame-by-frame disassembly and analysis, and recognizes the image data converted from the video stream data in an AI intelligent manner and generates the analysis result for application.
  • the server set up in the cloud performs the management, storage, and processing of video stream big data, and the processed image data is compared with the trained model including but not limited to brand classifiers, advertisement classifiers, and product classifiers in a distributed manner.
  • the analysis restore consumers' non-private real digital behaviors and produce more commercial value. In this process, there will be no delays, omissions, slow speed, large errors, high costs, and solve the bottleneck of consumer real-time digital behavior analysis. , Make business analysis closer to reality, bring more valuable analysis results to the brand, and guide the brand to optimize the consumption path globally.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Multimedia (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Medical Informatics (AREA)
  • Computing Systems (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Image Analysis (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种智能图像识别进行大数据采集分析***及应用方法，本发明的***包括智能云端服务器；智能云端服务器包括计算服务器和存储服务器，计算服务器中搭载有由数据读取模块、视频流数据处理模块、AI图像识别模块、数据存储模块、模型调优模块构成的图像识别***，存储服务器中设置有交互相连的视频流存储数据库、视频流管理模块、数据中心数据库。本发明还原消费者非隐私的真实数字行为，产出更多商业价值，这一过程中不产生延迟、遗漏、速度慢、误差大、成本高的问题，解决了消费者实时数字行为分析瓶颈，让商业分析更接近现实情况，给品牌方带来更多有价值的分析结果，去指导品牌全局优化消费路径。

Description

一种智能图像识别进行大数据采集分析***及应用方法
本申请要求于2019年9月23日提交中国专利局、申请号为201910898216.7、发明名称为“一种智能图像识别进行大数据采集分析***及应用方法”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及数据智能分析技术领域，具体地涉及一种智能图像识别进行大数据采集分析***及应用方法。
背景技术
随着社会的发展，目前针对用户数字行为数据的采集和分析，并进行规整与利用从而产生商业价值具有很大的应用前景，在企业进行产品或服务的商业市场推广时，需要掌握本类目的竞争对手产品或服务以及自身产品或服务的展现情况，包括品牌、产品、广告等具体的数字行为数据，供市场研究员进行分析。在用户数字行为数据产生的终端上，包括手机、平板、电脑、智能电视等，其产生的视频流数据若需要进行用户数字行为数据采集，现有的操作是采用人工进行数据处理，研究员将需要研究的且需要数据行为数据采集的视频交付，通过人工查找并标注的方式逐一的将该视频流中的含有包括品牌、产品、广告等具体的数字行为数据挑出来，并列表，一秒一秒的进行查看并标注，查找，数据易遗漏，且误差大，人工操作慢，成本高，且人工数字行为数据采集和分析的质量参差不齐，无法精准控制；相对于人工的数据采集和分析，计算机去分析品牌关心的和消费者关心的数据，并进行模块化的分析，具有人工无法达到的速度快、效率高、 精准、不易遗漏的优点。因此，针对以上问题提供一种智能图像识别进行大数据采集分析***及应用方法具有重要意义。
发明内容
本发明的目的在于提供一种智能图像识别进行大数据采集分析***及应用方法，通过数字行为转化成视频流的方式传输至智能云端服务器进行逐帧拆解分析，以AI智能的方式对视频流数据转化的图像数据进行识别并产生分析结果进行应用，这一过程中由设置于云端的服务器进行视频流大数据的管理、存储、处理并将处理后的图像数据与训练的包括但不限的品牌分类器、广告分类器、产品分类器模型进行分布式地比对和分析后，还原消费者非隐私的真实数字行为，产出更多商业价值，这一过程中不产生延迟、遗漏、速度慢、误差大、成本高的问题，解决了消费者实时数字行为分析瓶颈，让商业分析更接近现实情况，给品牌方带来更多有价值的分析结果，去指导品牌全局优化消费路径。
为解决上述技术问题，本发明是通过以下技术方案实现的：
本发明的一种智能图像识别进行大数据采集分析***，包括智能云端服务器，所述智能云端服务器包括交互相连的计算服务器和存储服务器，所述计算服务器中搭载有由依次相连的数据读取模块、视频流数据处理模块、AI图像识别模块、数据存储模块、模型调优模块构成的图像识别***，所述存储服务器中设置有交互相连的视频流存储数据库、视频流管理模块、数据中心数据库；所述数据读取模块与视频流存储数据库交互相连；所述智能云端服务器通过通讯网络和HTTP(S)协议接收经视频流采集终端采集的包括但不限于图片、直播视频、普通视频以视频流数据传送的大数据，并存储于视频流存储数据库；所述视频流存储数据库用于存储获取的视频流数据，所述数据中心数据库用于存储用户视频流采集终端上的设备信息，并将获取视频流与视频流采集终端相匹配；所述视频流管理模块用于管理视频流的上传、删除和读取顺序排列；所述数据读取模块用于按序读取视 频流存储数据库中已排列的视频流数据；所述视频流数据处理模块用于对视频流数据进行处理，筛选并加工成能够被AI图像识别模块有效识别的图像数据；所述AI图像识别模块用于对经视频数据处理模块处理后的图像数据进行AI智能图像识别并与存储于数据存储模块中的图像识别分类器模型进行比对获得识别结果数据；所述模型调优模块用于针对AI图像识别模块的召回率和正确率问题，通过追加训练分类器模型和分类样本对AI图像识别模块进行识别模型调优，使识别更加精准和细化。
进一步地，所述通讯网络采用包括但不仅限于4G、5G、WIFI网络。
进一步地，所述视频流采集终端采用能够对用户操作数字行为进行记录并形成视频流数据的任何终端，包括但不仅限于电脑、智能手机、平板电脑、智能电视。
进一步地，所述数据中心数据库内设置有设备视频流存储路径数据表、用户配置数据表、设备状态数据表；所述设备视频流存储路径数据表用于将获取的用户视频流数据与视频流存储数据库的分区间建立并分配存储路径；所述用户配置数据表用于获取和存储传输视频流数据的用户设备的配置参数数据信息；所述设备状态数据表用于实时获取并存储传输视频流数据的用户设备的设备状态数据信息，包括正在运行、中断、终止。
进一步地，所述数据存储模块中设置有包括但不仅限于广告分类器模型、品牌分类器模型、产品分类器模型。
一种智能图像识别进行大数据采集分析***的应用方法，包括如下步骤：
S01：获取视频流数据，视频流采集终端通过不限于录屏的方式获取用户在使用终端过程中产生的非隐私的安全视频流数据；S02：传输并存储视频流数据，通过通讯网络将视频流数据传输至智能云端 服务器中的视频流存储数据库中进行存储，同时匹配用户视频获取终端对应于视频流存储数据库中的存储路径、存储和记录用户视频获取终端的配置数据以及设备状态数据；S03：智能云端服务器进行数据读取，通过数据读取模块将视频流存储数据库中排序待读取的视频流数据进行读取并传至视频流数据处理模块进行处理；S04：智能云端服务器对经数据读取模块读取的视频流数据进行处理，包括两种处理：第一种，直接通过对获取的视频流数据进行视频关键帧的逐帧拆解后，对逐帧拆解后获得的连续时序的图像进行差异化分析和比对，选出并拼接出需要图像识别的关键帧发送至AI图像识别模块进行图像识别；第二种，对获取的视频流数据进行视频流自身特有的物理特征提取，与存储服务器中设定的检测模型进行差异化分析和比对，挑选出有用的视频流数据后，经第一种方式的视频关键帧进行逐帧拆解获取图像识别关键帧后发送至AI图像识别模块进行图像识别；由智能云端服务器根据读取的视频数据流的大小进行自动分配处理方式；S05：针对处理后的图像数据进行AI识别，根据预先训练的包括但不限于广告分类器、品牌分类器、产品分类器对图像进行分布式识别；若识别成功，反馈给智能云端服务器对应的参数的结果及对应类型；若识别失败，则放弃该数据；S06：对获取的经识别后的图像数据进行归类存储：针对反馈的数据进行再归类，得到完整的消费者的行为及接触的数据并存储于存储服务器中；S07：对AI图像识别模型进行调优：根据所述AI图像识别模块有召回率和正确率的问题，不断追加样本和测试对AI图像识别模型进行调优。
进一步地，步骤S04中的物理特征包括视频流数据的颜色特征、视频纹理特征、视频运动特性特征。
本发明具有以下有益效果：
本发明通过数字行为转化成视频流的方式传输至智能云端服务器进行逐帧拆解分析，以AI智能的方式对视频流数据转化的图像数据进行识别并产生分析结果进行应用，这一过程中由设置于云端的服 务器进行视频流大数据的管理、存储、处理并将处理后的图像数据与训练的包括但不限的品牌分类器、广告分类器、产品分类器模型进行分布式地比对和分析后，还原消费者非隐私的真实数字行为，产出更多商业价值，这一过程中不产生延迟、遗漏、速度慢、误差大、成本高的问题，解决了消费者实时数字行为分析瓶颈，让商业分析更接近现实情况，给品牌方带来更多有价值的分析结果，去指导品牌全局优化消费路径。
当然，实施本发明的任一产品并不一定需要同时达到以上所述的所有优点。
附图说明
图1为本发明的一种智能图像识别进行大数据采集分析***的计算服务器与存储服务器的连接与结构示意图；
图2为本发明的硬件连接结构示意图；
图3为图1中数据中心数据库的结构图；
图4为图1中数据存储模块的结构图；
图5为本发明的一种智能图像识别进行大数据采集分析***的应用方法的步骤图。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
请参阅图1-4所示，本发明的一种智能图像识别进行大数据采集分析***，包括智能云端服务器，智能云端服务器包括交互相连的计 算服务器和存储服务器，计算服务器中搭载有由依次相连的数据读取模块、视频流数据处理模块、AI图像识别模块、数据存储模块、模型调优模块构成的图像识别***，存储服务器中设置有交互相连的视频流存储数据库、视频流管理模块、数据中心数据库；数据读取模块与视频流存储数据库交互相连；
智能云端服务器通过通讯网络和HTTP(S)协议接收经视频流采集终端采集的包括但不限于图片、直播视频、普通视频以视频流数据传送的大数据，并存储于视频流存储数据库；
视频流存储数据库用于存储获取的视频流数据，数据中心数据库用于存储用户视频流采集终端上的设备信息，并将获取视频流与视频流采集终端相匹配；视频流管理模块用于管理视频流的上传、删除和读取顺序排列；
数据读取模块用于按序读取视频流存储数据库中已排列的视频流数据；视频流数据处理模块用于对视频流数据进行处理，筛选并加工成能够被AI图像识别模块有效识别的图像数据；AI图像识别模块用于对经视频数据处理模块处理后的图像数据进行AI智能图像识别并与存储于数据存储模块中的图像识别分类器模型进行比对获得识别结果数据；模型调优模块用于针对AI图像识别模块的召回率和正确率问题，通过追加训练分类器模型和分类样本对AI图像识别模块进行识别模型调优，使识别更加精准和细化。
其中，通讯网络采用包括但不仅限于4G、5G、WIFI网络。
其中，视频流采集终端采用能够对用户操作数字行为进行记录并形成视频流数据的任何终端，包括但不仅限于电脑、智能手机、平板电脑、智能电视。
其中，数据中心数据库内设置有设备视频流存储路径数据表、用户配置数据表、设备状态数据表；设备视频流存储路径数据表用于将获取的用户视频流数据与视频流存储数据库的分区间建立并分配存 储路径；用户配置数据表用于获取和存储传输视频流数据的用户设备的配置参数数据信息；设备状态数据表用于实时获取并存储传输视频流数据的用户设备的设备状态数据信息，包括正在运行、中断、终止。
其中，数据存储模块中设置有包括但不仅限于广告分类器模型、品牌分类器模型、产品分类器模型。
如图5所示，一种智能图像识别进行大数据采集分析***的应用方法，包括如下步骤：
S01：获取视频流数据，视频流采集终端通过不限于录屏的方式获取用户在使用终端过程中产生的非隐私的安全视频流数据；
S02：传输并存储视频流数据，通过通讯网络将视频流数据传输至智能云端服务器中的视频流存储数据库中进行存储，同时匹配用户视频获取终端对应于视频流存储数据库中的存储路径、存储和记录用户视频获取终端的配置数据以及设备状态数据；
S03：智能云端服务器进行数据读取，通过数据读取模块将视频流存储数据库中排序待读取的视频流数据进行读取并传至视频流数据处理模块进行处理；
S04：智能云端服务器对经数据读取模块读取的视频流数据进行处理，包括两种处理：
第一种，直接通过对获取的视频流数据进行视频关键帧的逐帧拆解后，对逐帧拆解后获得的连续时序的图像进行差异化分析和比对，选出并拼接出需要图像识别的关键帧发送至AI图像识别模块进行图像识别；
第二种，对获取的视频流数据进行视频流自身特有的物理特征提取，与存储服务器中设定的检测模型进行差异化分析和比对，挑选出有用的视频流数据后，经第一种方式的视频关键帧进行逐帧拆解获取图像识别关键帧后发送至AI图像识别模块进行图像识别；由智能云端服务器根据读取的视频数据流的大小进行自动分配处理方式；
S05：针对处理后的图像数据进行AI识别，根据预先训练的包括但不限于广告分类器、品牌分类器、产品分类器对图像进行分布式识别；若识别成功，反馈给智能云端服务器对应的参数的结果及对应类型；若识别失败，则放弃该数据；
S06：对获取的经识别后的图像数据进行归类存储：针对反馈的数据进行再归类，得到完整的消费者的行为及接触的数据并存储于存储服务器中；
S07：对AI图像识别模型进行调优：根据AI图像识别模块有召回率和正确率的问题，不断追加样本和测试对AI图像识别模型进行调优。
其中，步骤S04中的物理特征包括视频流数据的颜色特征、视频纹理特征、视频运动特性特征。
本发明通过数字行为转化成视频流的方式传输至智能云端服务器进行逐帧拆解分析，以AI智能的方式对视频流数据转化的图像数据进行识别并产生分析结果进行应用，这一过程中由设置于云端的服务器进行视频流大数据的管理、存储、处理并将处理后的图像数据与训练的包括但不限的品牌分类器、广告分类器、产品分类器模型进行分布式地比对和分析后，还原消费者非隐私的真实数字行为，产出更多商业价值，这一过程中不产生延迟、遗漏、速度慢、误差大、成本高的问题，解决了消费者实时数字行为分析瓶颈，让商业分析更接近现实情况，给品牌方带来更多有价值的分析结果，去指导品牌全局优化消费路径。
需要说明的是，本发明实施例描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
以上公开的本发明优选实施例只是用于帮助阐述本发明。优选实施例并没有详尽叙述所有的细节，也不限制该发明仅为所述的具体实 施方式。显然，根据本说明书的内容，可作很多的修改和变化。本说明书选取并具体描述这些实施例，是为了更好地解释本发明的原理和实际应用，从而使所属技术领域技术人员能很好地理解和利用本发明。本发明仅受权利要求书及其全部范围和等效物的限制。

Claims (7)

  1. 一种智能图像识别进行大数据采集分析***，包括智能云端服务器，其特征在于：
    所述智能云端服务器包括交互相连的计算服务器和存储服务器，所述计算服务器中搭载有由依次相连的数据读取模块、视频流数据处理模块、AI图像识别模块、数据存储模块、模型调优模块构成的图像识别***，所述存储服务器中设置有交互相连的视频流存储数据库、视频流管理模块、数据中心数据库；所述数据读取模块与视频流存储数据库交互相连；
    所述智能云端服务器通过通讯网络和HTTP(S)协议接收经视频流采集终端采集的包括但不限于图片、直播视频、普通视频以视频流数据传送的大数据，并存储于视频流存储数据库；
    所述视频流存储数据库用于存储获取的视频流数据，所述数据中心数据库用于存储用户视频流采集终端上的设备信息，并将获取视频流与视频流采集终端相匹配；所述视频流管理模块用于管理视频流的上传、删除和读取顺序排列；
    所述数据读取模块用于按序读取视频流存储数据库中已排列的视频流数据；所述视频流数据处理模块用于对视频流数据进行处理，筛选并加工成能够被AI图像识别模块有效识别的图像数据；所述AI图像识别模块用于对经视频数据处理模块处理后的图像数据进行AI智能图像识别并与存储于数据存储模块中的图像识别分类器模型进行比对获得识别结果数据；所述模型调优模块用于针对AI图像识别模块的召回率和正确率问题，通过追加训练分类器模型和分类样本对AI图像识别模块进行识别模型调优，使识别更加精准和细化。
  2. 根据权利要求1所述的一种智能图像识别进行大数据采集分析***，其特征在于，所述通讯网络包括但不仅限于4G、5G、WIFI 网络。
  3. 根据权利要求1所述的一种智能图像识别进行大数据采集分析***，其特征在于，所述视频流采集终端采用能够对用户操作数字行为进行记录并形成视频流数据的任何终端，包括但不仅限于电脑、智能手机、平板电脑、智能电视。
  4. 根据权利要求1所述的一种智能图像识别进行大数据采集分析***，其特征在于，所述数据中心数据库内设置有设备视频流存储路径数据表、用户配置数据表、设备状态数据表；
    所述设备视频流存储路径数据表用于将获取的用户视频流数据与视频流存储数据库的分区间建立并分配存储路径；
    所述用户配置数据表用于获取和存储传输视频流数据的用户设备的配置参数数据信息；
    所述设备状态数据表用于实时获取并存储传输视频流数据的用户设备的设备状态数据信息，包括正在运行、中断、终止。
  5. 根据权利要求1所述的一种智能图像识别进行大数据采集分析***，其特征在于，所述数据存储模块中设置有包括但不仅限于广告分类器模型、品牌分类器模型、产品分类器模型。
  6. 根据权利要求1-5任一项所述的一种智能图像识别进行大数据采集分析***的应用方法，其特征在于，包括如下步骤：
    S01：获取视频流数据，视频流采集终端通过不限于录屏的方式获取用户在使用终端过程中产生的非隐私的安全视频流数据；
    S02：传输并存储视频流数据，通过通讯网络将视频流数据传输至智能云端服务器中的视频流存储数据库中进行存储，同时匹配用户视频获取终端对应于视频流存储数据库中的存储路径、存储和记录用户视频获取终端的配置数据以及设备状态数据；
    S03：智能云端服务器进行数据读取，通过数据读取模块将视频流 存储数据库中排序待读取的视频流数据进行读取并传至视频流数据处理模块进行处理；
    S04：智能云端服务器对经数据读取模块读取的视频流数据进行处理，包括两种处理：
    第一种，直接通过对获取的视频流数据进行视频关键帧的逐帧拆解后，对逐帧拆解后获得的连续时序的图像进行差异化分析和比对，选出并拼接出需要图像识别的关键帧发送至AI图像识别模块进行图像识别；
    第二种，对获取的视频流数据进行视频流自身特有的物理特征提取，与存储服务器中设定的检测模型进行差异化分析和比对，挑选出有用的视频流数据后，经第一种方式的视频关键帧进行逐帧拆解获取图像识别关键帧后发送至AI图像识别模块进行图像识别；由智能云端服务器根据读取的视频数据流的大小进行自动分配处理方式；
    S05：针对处理后的图像数据进行AI识别，根据预先训练的包括但不限于广告分类器、品牌分类器、产品分类器对图像进行分布式识别；若识别成功，反馈给智能云端服务器对应的参数的结果及对应类型；若识别失败，则放弃该数据；
    S06：对获取的经识别后的图像数据进行归类存储：针对反馈的数据进行再归类，得到完整的消费者的行为及接触的数据并存储于存储服务器中；
    S07：对AI图像识别模型进行调优：根据所述AI图像识别模块有召回率和正确率的问题，不断追加样本和测试对AI图像识别模型进行调优。
  7. 根据权利要求6所述的一种智能图像识别进行大数据采集分析***的应用方法，其特征在于，步骤S04中的物理特征包括视频流数据的颜色特征、视频纹理特征、视频运动特性特征。
PCT/CN2020/102654 2019-09-23 2020-07-17 一种智能图像识别进行大数据采集分析***及应用方法 WO2021057213A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US17/294,105 US11483621B2 (en) 2019-09-23 2020-07-17 Big data acquisition and analysis system using intelligent image recognition, and application method thereof
EP20867413.5A EP3866057A4 (en) 2019-09-23 2020-07-17 BIG DATA ACQUISITION AND ANALYSIS SYSTEM BASED ON INTELLIGENT IMAGE RECOGNITION, AND METHOD OF APPLICATION

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910898216.7A CN110826398B (zh) 2019-09-23 2019-09-23 一种智能图像识别进行大数据采集分析***及应用方法
CN201910898216.7 2019-09-23

Publications (1)

Publication Number Publication Date
WO2021057213A1 true WO2021057213A1 (zh) 2021-04-01

Family

ID=69548267

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/102654 WO2021057213A1 (zh) 2019-09-23 2020-07-17 一种智能图像识别进行大数据采集分析***及应用方法

Country Status (4)

Country Link
US (1) US11483621B2 (zh)
EP (1) EP3866057A4 (zh)
CN (1) CN110826398B (zh)
WO (1) WO2021057213A1 (zh)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113313033A (zh) * 2021-05-31 2021-08-27 常州汉腾自动化设备有限公司 一种基于物联网传感器的供热管网数字监控***
CN113672425A (zh) * 2021-08-20 2021-11-19 浙江创意声光电科技有限公司 一种智慧照明***的综合运行态势实时分析***
CN114827400A (zh) * 2022-03-23 2022-07-29 南京华脉科技股份有限公司 一种基于区块链大数据的图像处理***
CN115190320A (zh) * 2022-06-09 2022-10-14 武汉鸿宸信信息科技有限公司 一种用于视频处理与存储的云计算***和方法
CN115242664A (zh) * 2022-06-17 2022-10-25 江苏电力信息技术有限公司 基于大数据分析模型的机房智能管理方法
CN116360992A (zh) * 2023-03-30 2023-06-30 郑州地铁集团有限公司运营分公司 基于容器化微服务的轨道交通供电智能运维方法及***
CN117764303A (zh) * 2023-11-17 2024-03-26 南京公路发展(集团)有限公司 一种基于人工智能的道路巡检数据分析***及方法
CN118069894A (zh) * 2024-04-12 2024-05-24 乾健科技有限公司 一种大数据存储管理方法及***

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110826398B (zh) * 2019-09-23 2021-04-02 上海意略明数字科技股份有限公司 一种智能图像识别进行大数据采集分析***及应用方法
CN111415014B (zh) * 2020-03-20 2023-10-20 中国建设银行股份有限公司 一种模型预测的结果数据管理***及方法
CN111857914B (zh) * 2020-07-06 2021-10-01 上海意略明数字科技股份有限公司 终端上的应用程序操作信息获取方法及装置、存储介质、终端
CN112287158A (zh) * 2020-11-05 2021-01-29 广州增强信息科技有限公司 用于包装上数字化防伪、数字化内容展示与数据分析的***与方法
CN112702430B (zh) * 2020-12-24 2022-07-29 山东大学 一种基于云边模式和Web技术的数据传输方法
CN112887759A (zh) * 2021-01-22 2021-06-01 Oppo广东移动通信有限公司 智能电视的信号处理方法、装置、智能电视以及存储介质
CN113342991B (zh) * 2021-06-08 2021-12-21 南京智云兴禾信息科技有限公司 一种基于人工智能的知识图谱融合***及其方法
CN113326849B (zh) * 2021-07-20 2022-01-11 广东魅视科技股份有限公司 一种可视化数据采集方法及***
CN114401254B (zh) 2022-03-22 2022-11-25 北京百度网讯科技有限公司 流媒体服务的处理方法、装置、电子设备及存储介质
CN114826483B (zh) * 2022-03-28 2022-11-15 慧之安信息技术股份有限公司 基于边缘计算的智能网络传输计算***和方法
CN114881808B (zh) * 2022-04-01 2024-01-23 云南电网有限责任公司昭通供电局 基于大数据的电力窃电精准识别方法及防窃电***
CN115952314A (zh) * 2023-03-15 2023-04-11 潍坊职业学院 一种基于数据识别的数据处理***
CN116957291B (zh) * 2023-08-10 2024-05-07 湛江市广厦施工图审查服务中心 一种工程管理数据智能管理***及方法
CN117291291B (zh) * 2023-08-12 2024-04-23 江苏信实环境工程有限公司 一种基于物联网的虫情智能监控***及方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160004914A1 (en) * 2014-07-01 2016-01-07 Itx Security Co., Ltd. Intelligent video analysis system and method
CN109271554A (zh) * 2018-09-07 2019-01-25 江西省云眼大视界科技有限公司 一种智能视频识别***及其应用
CN110062208A (zh) * 2019-04-23 2019-07-26 上海赫煊自动化***工程有限公司 一种安防智能实时分析***和方法
CN110163154A (zh) * 2019-05-23 2019-08-23 湖南机电职业技术学院 基于人工智能的视频监控***
CN110826398A (zh) * 2019-09-23 2020-02-21 上海意略明数字科技股份有限公司 一种智能图像识别进行大数据采集分析***及应用方法

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8059865B2 (en) 2007-11-09 2011-11-15 The Nielsen Company (Us), Llc Methods and apparatus to specify regions of interest in video frames
US8948515B2 (en) * 2010-03-08 2015-02-03 Sightera Technologies Ltd. Method and system for classifying one or more images
CN101799876B (zh) * 2010-04-20 2011-12-14 王巍 一种视音频智能分析管控***
US8873813B2 (en) * 2012-09-17 2014-10-28 Z Advanced Computing, Inc. Application of Z-webs and Z-factors to analytics, search engine, learning, recognition, natural language, and other utilities
US11914674B2 (en) * 2011-09-24 2024-02-27 Z Advanced Computing, Inc. System and method for extremely efficient image and pattern recognition and artificial intelligence platform
US9135712B2 (en) 2012-08-01 2015-09-15 Augmented Reality Lab LLC Image recognition system in a cloud environment
US10275488B1 (en) * 2014-12-09 2019-04-30 Cloud & Stream Gears Llc Incremental covariance calculation for big data or streamed data using components
US10390082B2 (en) 2016-04-01 2019-08-20 Oath Inc. Computerized system and method for automatically detecting and rendering highlights from streaming videos
US10482345B2 (en) * 2016-06-23 2019-11-19 Capital One Services, Llc Systems and methods for automated object recognition
US10986152B2 (en) * 2016-12-29 2021-04-20 Arris Enterprises Llc Method for dynamically managing content delivery
CN108334518A (zh) * 2017-01-20 2018-07-27 ***通信有限公司研究院 一种广告加载方法及装置
US20190122121A1 (en) * 2017-10-23 2019-04-25 AISA Innotech Inc. Method and system for generating individual microdata
CN108833933A (zh) * 2018-06-14 2018-11-16 广东互通宽带网络服务有限公司 一种使用支持向量机推荐视频流量的方法及***
CN110149530B (zh) * 2018-06-15 2021-08-24 腾讯科技（深圳）有限公司 一种视频处理方法和装置
US20200007934A1 (en) * 2018-06-29 2020-01-02 Advocates, Inc. Machine-learning based systems and methods for analyzing and distributing multimedia content
CN109885562A (zh) * 2019-01-17 2019-06-14 安徽谛听信息科技有限公司 一种基于网络空间安全的大数据智能分析***
US20200364583A1 (en) * 2019-05-14 2020-11-19 Robert D. Pedersen Iot sensor network artificial intelligence warning, control and monitoring systems and methods

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160004914A1 (en) * 2014-07-01 2016-01-07 Itx Security Co., Ltd. Intelligent video analysis system and method
CN109271554A (zh) * 2018-09-07 2019-01-25 江西省云眼大视界科技有限公司 一种智能视频识别***及其应用
CN110062208A (zh) * 2019-04-23 2019-07-26 上海赫煊自动化***工程有限公司 一种安防智能实时分析***和方法
CN110163154A (zh) * 2019-05-23 2019-08-23 湖南机电职业技术学院 基于人工智能的视频监控***
CN110826398A (zh) * 2019-09-23 2020-02-21 上海意略明数字科技股份有限公司 一种智能图像识别进行大数据采集分析***及应用方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3866057A4 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113313033A (zh) * 2021-05-31 2021-08-27 常州汉腾自动化设备有限公司 一种基于物联网传感器的供热管网数字监控***
CN113313033B (zh) * 2021-05-31 2023-12-05 常州汉腾自动化设备有限公司 一种基于物联网传感器的供热管网数字监控***
CN113672425A (zh) * 2021-08-20 2021-11-19 浙江创意声光电科技有限公司 一种智慧照明***的综合运行态势实时分析***
CN114827400A (zh) * 2022-03-23 2022-07-29 南京华脉科技股份有限公司 一种基于区块链大数据的图像处理***
CN114827400B (zh) * 2022-03-23 2024-05-10 趣橙(上海)信息技术有限公司 一种基于区块链大数据的图像处理***
CN115190320A (zh) * 2022-06-09 2022-10-14 武汉鸿宸信信息科技有限公司 一种用于视频处理与存储的云计算***和方法
CN115242664A (zh) * 2022-06-17 2022-10-25 江苏电力信息技术有限公司 基于大数据分析模型的机房智能管理方法
CN116360992A (zh) * 2023-03-30 2023-06-30 郑州地铁集团有限公司运营分公司 基于容器化微服务的轨道交通供电智能运维方法及***
CN116360992B (zh) * 2023-03-30 2023-11-17 郑州地铁集团有限公司运营分公司 基于容器化微服务的轨道交通供电智能运维方法及***
CN117764303A (zh) * 2023-11-17 2024-03-26 南京公路发展(集团)有限公司 一种基于人工智能的道路巡检数据分析***及方法
CN118069894A (zh) * 2024-04-12 2024-05-24 乾健科技有限公司 一种大数据存储管理方法及***

Also Published As

Publication number Publication date
CN110826398A (zh) 2020-02-21
CN110826398B (zh) 2021-04-02
EP3866057A1 (en) 2021-08-18
EP3866057A4 (en) 2022-07-20
US11483621B2 (en) 2022-10-25
US20220030310A1 (en) 2022-01-27

Similar Documents

Publication Publication Date Title
WO2021057213A1 (zh) 一种智能图像识别进行大数据采集分析***及应用方法
WO2021238631A1 (zh) 物品信息的显示方法、装置、设备及可读存储介质
US20210256320A1 (en) Machine learning artificialintelligence system for identifying vehicles
CN112565825B (zh) 一种视频数据处理方法、装置、设备以及介质
US20140074866A1 (en) System and method for enhancing metadata in a video processing environment
WO2017035519A1 (en) Supervised learning based recommendation system
CN112765373B (zh) 资源推荐方法、装置、电子设备和存储介质
WO2017080173A1 (zh) 基于自然信息识别的推送***和方法及一种客户端
CN111432282B (zh) 一种视频推荐方法及装置
CN103020172B (zh) 一种利用视频信息搜索物品的方法和装置
CN105323601A (zh) 基于多屏用户行为数据的人员属性标识方法
WO2022042157A1 (zh) 视频数据的制作方法、装置、计算机设备和存储介质
EP3707612B1 (en) Duplicative data detection
EP3414680A1 (en) Text digest generation for searching multiple video streams
CN113268645A (zh) 信息召回方法、模型训练方法、装置、设备及存储介质
CN112052387A (zh) 一种内容推荐方法、装置和计算机可读存储介质
WO2023231542A1 (zh) 表示信息的确定方法、装置、设备及存储介质
Tao et al. Real-time personalized content catering via viewer sentiment feedback: a QoE perspective
CN113992973B (zh) 视频摘要生成方法、装置、电子设备和存储介质
CN116051192A (zh) 处理数据的方法和装置
CN103220555B (zh) 一种数字电视用户的分类方法、装置及***
US20230334314A1 (en) Content recommendation method and apparatus, device, storage medium, and program product
CN115171014A (zh) 视频处理方法、装置、电子设备及计算机可读存储介质
Wang et al. CoLive: Edge-Assisted Clustered Learning Framework for Viewport Prediction in 360$^\circ $ Live Streaming
CN111143688A (zh) 一种基于移动新闻客户端的评估方法及其***

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20867413

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020867413

Country of ref document: EP

Effective date: 20210512

NENP Non-entry into the national phase

Ref country code: DE