WO2019062388A1 - 广告效果分析方法及装置 - Google Patents

广告效果分析方法及装置 Download PDF

Info

Publication number
WO2019062388A1
WO2019062388A1 PCT/CN2018/101634 CN2018101634W WO2019062388A1 WO 2019062388 A1 WO2019062388 A1 WO 2019062388A1 CN 2018101634 W CN2018101634 W CN 2018101634W WO 2019062388 A1 WO2019062388 A1 WO 2019062388A1
Authority
WO
WIPO (PCT)
Prior art keywords
identified
video
target
preset model
advertisement
Prior art date
Application number
PCT/CN2018/101634
Other languages
English (en)
French (fr)
Inventor
王天祎
戴威
Original Assignee
北京国双科技有限公司
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 北京国双科技有限公司 filed Critical 北京国双科技有限公司
Publication of WO2019062388A1 publication Critical patent/WO2019062388A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0242—Determining effectiveness of advertisements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising

Definitions

  • the embodiments of the present invention relate to the field of new media technologies, and in particular, to an advertisement effect analysis method and apparatus.
  • the present application provides an advertisement effect analysis method and apparatus for achieving the purpose of quantitative evaluation and analysis of advertisement effects in existing video streaming media.
  • a first aspect of the present invention provides an advertising effect analysis method, including:
  • the calculation results are summarized and counted to obtain an advertisement effect analysis result.
  • the preset model is generated by the target creative training, including:
  • Obtaining a target creative from a target platform including a search platform and/or a material website, based on a crawler technology
  • the target creative is batch-labeled to generate a corresponding training set, and the target creative includes a brand product image, a brand logo, a brand advertisement material, or a combination of any elements;
  • the training set is trained to obtain a corresponding preset model.
  • the training is performed on the training set to obtain a corresponding preset model, including:
  • the training set corresponding to each element is trained based on the tensorflow and the Faster-RCNN architecture to obtain a corresponding preset model.
  • the valid data includes: display duration, number of times, and area data, and calculate valid data when each of the elements is displayed in the to-be-identified video, and obtain corresponding calculation results, including:
  • the to-be-identified video includes an offline video, a live video, and/or a dynamic video map.
  • a second aspect of the present invention provides an advertisement effect analysis apparatus, including:
  • a disassembling unit configured to disassemble a video to be identified by a frame, and obtain a set of images to be identified
  • An annotation unit configured to identify the to-be-identified image collection based on a predetermined preset model, determine an element related to the target creative material, and mark, the preset model is trained by the target creative material;
  • a calculating unit configured to calculate valid data of each of the determined elements when displayed in the to-be-identified video, to obtain a corresponding calculation result
  • An analysis unit is configured to summarize and calculate the calculation result to obtain an advertisement effect analysis result.
  • the device further comprises:
  • a search unit configured to acquire a target creative from a target platform, including a search platform and/or a material website, based on a crawler technology
  • a batch annotation unit configured to batch mark the target creative, and generate a corresponding training set, where the target creative includes a brand product image, a brand logo, a brand advertisement material, or a combination of any elements;
  • the training unit is configured to train the training set to obtain a corresponding preset model.
  • the valid data includes: a display duration, a number of times, and an area data
  • the calculating unit 503 is configured to calculate a display duration, a number of times, and an area data of each of the determined elements in the to-be-identified video, The obtained display duration, the number of times, and the area data are obtained as calculation results.
  • a storage medium comprising a stored program, wherein the device in which the storage medium is located is controlled to execute an advertisement effect analysis method provided by the first aspect of the present invention when the program is running.
  • a processor for running a program wherein the program runtime executes the advertisement effect analysis method provided by the first aspect of the present invention.
  • the present invention discloses an advertisement effect analysis method and apparatus. Obtaining a to-be-identified image set by disassembling the to-be-identified video by frame, and then identifying the to-be-identified image collection based on a predetermined preset model, determining an element related to the target creative material, and marking the preset
  • the model is generated by the target creative training; the valid data of each determined element is displayed in the to-be-identified video, and the corresponding calculation result is obtained; finally, the calculation result is summarized and counted, and the result is obtained.
  • the invention identifies the element related to the target creative material in the to-be-identified video, and calculates valid data when the element is displayed in the to-be-identified video, performs statistics and analysis based on the valid data, and obtains an analysis result reflecting the advertisement effect, and realizes Quantitative evaluation and analysis of the effectiveness of advertising in existing video streaming media. Further, the company is able to understand the advertising effects that its corporate brand can bring when displayed in the video.
  • FIG. 1 is a schematic flowchart diagram of an advertisement effect analysis method according to an embodiment of the present invention
  • FIG. 2 is a manner of displaying an element related to a target creative material in a video according to an embodiment of the present invention
  • FIG. 3 is a diagram showing a result of batch labeling according to an embodiment of the present invention.
  • FIG. 4 is a display diagram of an advertisement effect analysis result disclosed in an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of an advertisement effect analysis apparatus according to an embodiment of the present invention.
  • the present invention discloses an advertisement effect analysis method for achieving the purpose of evaluating or analyzing an advertisement effect in a video streaming media.
  • FIG. 1 is a schematic flowchart diagram of an advertisement effect analysis method according to an embodiment of the present invention.
  • each video is analyzed in the following manner.
  • the to-be-identified video includes offline video, live video, and/or dynamic video map.
  • Step S101 Disassemble the to-be-identified video by frame, and obtain a to-be-identified image set.
  • a video to be identified is determined, and when the step S101 is performed, the to-be-identified video is disassembled in a frame by frame. Each frame of the image obtained by the set is disassembled to obtain a set of images to be identified.
  • Step S102 Identify the to-be-identified image set based on a predetermined preset model, determine an element related to the target creative material, and mark it.
  • the preset model is generated by the target creative training.
  • the specific process is shown in Figure 2:
  • Step S201 Obtain a target creative material from the target platform based on the crawler technology.
  • the target platform includes a search platform and/or a material website.
  • the target creative includes a brand product image, a brand logo, any one of the branded advertising materials, or a combination of any of the elements.
  • step S201 based on the crawler technology, the brand product image, the brand logo logo, and the brand advertisement material that are required to be analyzed and evaluated are acquired from the search platform for batch collection and storage.
  • Step S202 batch labeling the target creative material to generate a corresponding training set.
  • the crowdsourcing system or other annotation tools may be used to batch label each element included in the target creative to generate a training set corresponding to each element. It is also possible to automatically batch-label each element included in the target brand material by using an automatic annotation tool generated by the learning to generate a training set corresponding to each element.
  • the elements related to the Lingdu car brand need to be marked, as shown in Fig. 3, by using the words of Lingdu car and Lingdu car logo in the picture. Label.
  • Step S203 Train the training set to obtain a corresponding preset model.
  • the training set corresponding to each element may be trained based on the tensorflow and the Faster-RCNN architecture to obtain a corresponding preset model.
  • tensorflow is an artificial intelligence learning system, which is mainly used in many machine deep learning fields such as speech recognition or image recognition.
  • the Faster-RCNN architecture is used to achieve target detection.
  • the principle of target detection is as follows: firstly, the image is normalized, the candidate region is extracted therefrom, and then the feature is extracted from the extracted candidate region using the depth network, and the position of the candidate frame is adjusted.
  • features such as inception and Resnet can be used for feature extraction, so that the training effect of the model is more precise.
  • the preset model trained in the above manner can adaptively identify each element in the image related to the target creative.
  • Step S103 Calculate valid data when each of the labeled elements is displayed in the to-be-identified video, and obtain a corresponding calculation result.
  • the valid data includes: display duration, number of times, area data, and the like.
  • only one of the display duration, the number of displays, and the area may be calculated. The more valid data is calculated, the more the calculated results are, the more favorable the analysis of subsequent advertising effects is.
  • the valid data for each element is displayed when it is displayed in the video to be recognized. For example, calculating the display duration, the number of times, and the area data of each of the determined elements in the to-be-identified video, and using the obtained display duration, the number of times, and the area data as a calculation result.
  • Step S104 Summarize and count the calculation results to obtain an advertisement effect analysis result.
  • the above steps S101 to S104 are processes for analyzing a video. If it is necessary to analyze the advertisement effects in all the videos of the same program, all the videos may be analyzed by using the above steps S101 to S104.
  • An advertisement effect analysis method disclosed in the embodiment of the present invention obtains a to-be-identified image set by disassembling a to-be-identified video according to a frame, and then identifying the to-be-identified image set based on a predetermined preset model, determining and a target An element related to the creative, and the preset model is trained by the target creative; and calculating valid data of each of the determined elements in the to-be-identified video, and obtaining a corresponding calculation result; Finally, the calculation results are summarized and counted to obtain an advertisement effect analysis result.
  • the invention identifies the element related to the target creative material in the to-be-identified video, and calculates valid data when the element is displayed in the to-be-identified video, performs statistics and analysis based on the valid data, and obtains an analysis result reflecting the advertisement effect, and realizes Quantitative evaluation and analysis of the effectiveness of advertising in existing video streaming media. Further, the company is able to understand the advertising effects that its corporate brand can bring when displayed in the video.
  • the embodiment of the present invention further discloses an advertisement effect analysis device.
  • the advertisement effect analysis device 500 mainly includes:
  • the disassembling unit 501 is configured to disassemble the to-be-identified video by frame, and obtain a to-be-identified image set.
  • the labeling unit 502 is configured to identify the to-be-identified image set based on a predetermined preset model, determine an element related to the target creative material, and mark the preset model to be generated by the target creative material.
  • the calculating unit 503 is configured to calculate valid data when each of the labeled elements is displayed in the to-be-identified video, and obtain a corresponding calculation result.
  • the valid data includes: a display duration, a number of times, and an area data
  • the calculating unit 503 is configured to calculate a display duration, a number of times, and an area data of each of the determined elements in the to-be-identified video, The obtained display duration, the number of times, and the area data are obtained as calculation results.
  • the analyzing unit 504 is configured to summarize and calculate the calculation result to obtain an advertisement effect analysis result.
  • the advertisement effect analysis device further includes: a preset unit 505.
  • the preset unit 505 includes:
  • a search unit for obtaining a target creative from a target platform based on a crawler technology, the target platform including a search platform and/or a material website.
  • the batch annotation unit is configured to batch mark the target creative, and generate a corresponding training set, where the target creative includes a brand product image, a brand logo, a brand advertisement material, or a combination of any elements.
  • the batch annotation unit may use a crowdsourcing system or a labelimg tool to batch mark the elements included in the target creative to generate a training set corresponding to each element.
  • the batch annotation unit may also automatically batch label each element included in the target brand material by using an automatic annotation tool generated by the learning to generate a training set corresponding to each element.
  • the training unit is configured to train the training set to obtain a corresponding preset model.
  • the training unit may train the training set corresponding to each element based on the tensorflow and the Faster-RCNN architecture to obtain a corresponding preset model.
  • each unit in the advertisement effect analysis device disclosed in the embodiment of the present invention is the same as the advertisement effect analysis method disclosed in the above embodiment of the present invention, and can be referred to the advertisement effect analysis method disclosed in the above embodiment of the present invention.
  • the corresponding parts are not described here.
  • each of the above units may be implemented by a hardware device composed of a processor and a memory. Specifically, each of the above units and modules is stored in the memory as a program unit, and the program unit stored in the memory is executed by the processor to implement an analysis of the advertisement effect.
  • the processor includes a kernel, and the kernel removes the corresponding program unit from the memory.
  • the kernel can be set to one or more, and the analysis of the effects of the advertisement can be achieved by adjusting the kernel parameters.
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory (flash RAM), the memory including at least one Memory chip.
  • RAM random access memory
  • ROM read only memory
  • flash RAM flash memory
  • an embodiment of the present invention provides a processor, where the processor is configured to run a program, where the program is executed to execute the advertisement effect analysis method.
  • an embodiment of the present invention provides a device, including a processor, a memory, and a program stored on the memory and operable on the processor.
  • the processor executes the program, the following steps are implemented: disassembling the to-be-identified video by frame Obtaining a set of to-be-identified images; identifying the to-be-identified image set based on a predetermined preset model, determining an element related to the target creative material, and marking, the preset model being trained by the target creative material; Calculating valid data when each of the labeled elements is displayed in the to-be-identified video, and obtaining a corresponding calculation result; summarizing and counting the calculation result to obtain an advertisement effect analysis result.
  • the to-be-identified video includes an offline video, a live video, and/or a dynamic video map.
  • the valid data includes: display duration, number of times, and area data.
  • the process of training the preset model by the target creative includes: acquiring a target creative from a target platform based on a crawler technology, the target platform includes a search platform and/or a material website; and labeling the target advertisement in batches
  • the material generates a corresponding training set
  • the target creative includes a brand product image, a brand logo, a combination of any one of the brand advertisement materials or any of the elements; training the training set to obtain a corresponding pre- Set the model.
  • the crowdsourcing system or the labelimg tool may be used to batch-label the elements included in the target creative to generate a training set corresponding to each element.
  • the automatic annotation tool generated by the learning device may be used to automatically batch-label each element included in the target brand material to generate a training set corresponding to each element.
  • the training set corresponding to each element may be trained based on the tensorflow and the Faster-RCNN architecture to obtain a corresponding preset model.
  • the device disclosed in the embodiment of the present invention may be a server, a PC, a PAD, a mobile phone, or the like.
  • an embodiment of the present invention further provides a storage medium, where a program is stored, and the program implements the advertisement effect analysis method when executed by a processor.
  • the present application also provides a computer program product, when executed on a data processing device, adapted to perform a process of initializing a method of disassembling a video to be identified by a frame to obtain a set of images to be identified; based on a predetermined pre-determination Setting a model to identify the to-be-identified image set, determining an element related to the target creative material, and marking, the preset model is trained by the target creative material; calculating each of the labeled elements in the to-be-identified
  • the valid data displayed in the video is obtained, and the corresponding calculation result is obtained; the calculation result is summarized and counted, and the result of the advertisement effect analysis is obtained.
  • the to-be-identified video includes an offline video, a live video, and/or a dynamic video map.
  • the valid data includes: display duration, number of times, and area data.
  • the process of training the preset model by the target creative includes: acquiring a target creative from a target platform based on a crawler technology, the target platform includes a search platform and/or a material website; and labeling the target advertisement in batches
  • the material generates a corresponding training set
  • the target creative includes a brand product image, a brand logo, a combination of any one of the brand advertisement materials or any of the elements; training the training set to obtain a corresponding pre- Set the model.
  • the crowdsourcing system or the labelimg tool may be used to batch-label the elements included in the target creative to generate a training set corresponding to each element.
  • the training set corresponding to each element may be trained based on the tensorflow and the Faster-RCNN architecture to obtain a corresponding preset model.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media includes both permanent and non-persistent, removable and non-removable media.
  • Information storage can be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include temporary storage of computer readable media, such as modulated data signals and carrier waves.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Image Analysis (AREA)

Abstract

本发明公开了一种广告效果分析方法及装置，通过按帧拆解待识别视频，获取待识别图像集合，基于预先确定的预设模型对待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注；再计算批注的各个元素在待识别视频中显示时的有效数据，得到对应的计算结果；对计算结果进行汇总和统计，得到广告效果分析结果。本发明通过识别待识别视频中与目标广告素材相关的元素，并计算确定该元素在待识别视频中显示时的有效数据，基于该有效数据进行统计和分析，得到体现广告效果的分析结果，实现对现有视频流媒体中的广告效果进行量化评估和分析的目的。进一步，使得企业能够了解到自己的企业品牌在视频中显示时所能够带来的广告效果。

Description

广告效果分析方法及装置
本申请要求于2017年9月30日提交中国专利局、申请号为201710918855.6、申请名称为“广告效果分析方法及装置”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
本发明实施例涉及新媒体技术领域，更具体地说涉及一种广告效果分析方法及装置。
背景技术
近年来，越来越多的企业通过对电视节目和网络节目的冠名，或者对节目做出赞助，从而在节目中嵌入自己的广告，来增加企业品牌的曝光度。例如，加多宝、大众凌渡汽车在收视率较高的娱乐节目上都有做过冠名。
但是，虽然企业做了冠名或者赞助，对于企业品牌在节目中的曝光度所带来的广告效果如何，目前并没有一种有效的方式可以对视频流媒体中的广告效果进行评估或分析。
因此，目前亟需一种能够对视频流媒体中的广告效果进行评估或分析的方案。
发明内容
有鉴于此，本申请提供了一种广告效果分析方法及装置，以实现对现有视频流媒体中的广告效果进行量化评估和分析的目的。
为了实现上述目的，现提出的方案如下：
本发明第一方面提供了一种广告效果分析方法，包括：
按帧拆解待识别视频，获取待识别图像集合；
基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；
计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；
对所述计算结果进行汇总和统计，得到广告效果分析结果。
优选的，所述预设模型由所述目标广告素材训练生成，包括：
基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；
批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；
对所述训练集进行训练，得到对应的预设模型。
优选的，所述对所述训练集进行训练，得到对应的预设模型，包括：
基于tensorflow和Faster-RCNN架构对对应各个元素的所述训练集进行训练，得到对应的预设模型。
优选的，所述有效数据包括：显示时长、次数和面积数据，计算确定的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果，包括：
计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
优选的，所述待识别视频包括离线视频，直播视频和/或动态视频图。
本发明第二方面提供了一种广告效果分析装置，包括：
拆解单元，用于按帧拆解待识别视频，获取待识别图像集合；
标注单元，用于基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；
计算单元，用于计算确定的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；
分析单元，用于对所述计算结果进行汇总和统计，得到广告效果分析结果。
优选的，所述装置还包括：
搜索单元，用于基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；
批量批注单元，用于批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；
训练单元，用于对所述训练集进行训练，得到对应的预设模型。
可选的，所述有效数据包括：显示时长、次数和面积数据，所述计算单元503，用于计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
一种存储介质，所述存储介质包括存储的程序，其中，在所述程序运行时控制所述存储介质所在设备执行如本发明第一方面提供的广告效果分析方法。
一种处理器，所述处理器用于运行程序，其中，所述程序运行时执行本发明第一方面提供的广告效果分析方法。
经由上述技术方案可知，本发明公开一种广告效果分析方法及装置。通过按帧拆解待识别视频，获取待识别图像集合，然后，基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；再计算确定的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；最后，对所述计算结果进行汇总和统计，得到广告效果分析结果。本发明通过识别待识别视频中与目标广告素材相关的元素，并计算确定该元素在待识别视频中显示时的有效数据，基于该有效数据进行统计和分析，得到体现广告效果的分析结果，实现对现有视频流媒体中的广告效果进行量化评估和分析的目的。进一步，使得企业能够了解到自己的企业品牌在视频中显示时所能够带来的广告效果。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的实施例，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，还可以根据提供的附图获得其他的附图。
图1为本发明实施例公开的一种广告效果分析方法的流程示意图；
图2为本发明实施例公开的一种与目标广告素材相关的元素在视频中的显示方式；
图3为本发明实施例公开的一种批量标注的结果显示图；
图4为本发明实施例公开的广告效果分析结果显示图；
图5为本发明实施例公开的一种广告效果分析装置的结构示意图。
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做出创造性劳动前提下所获得的所有其他实 施例，都属于本发明保护的范围。
由背景技术可知，目前并没有一种有效的方式可以对视频流媒体中的广告效果进行评估或分析。因此，本发明公开了一种广告效果分析方法，以实现对视频流媒体中的广告效果进行评估或分析的目的。
如图1所示，为本发明实施例公开的一种广告效果分析方法的流程示意图。
在具体实现过程中，若需要识别的视频很多，则采用下述方式对每一个视频进行分析。该待识别视频包括离线视频，直播视频和/或动态视频图。
步骤S101：按帧拆解待识别视频，获取待识别图像集合。
确定一个待识别视频，在执行步骤S101时，将该待识别视频按照一帧一帧的方式进行拆解。集合拆解得到的每一帧图像，得到待识别图像集合。
步骤S102：基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注。
在具体实现中，所述预设模型由所述目标广告素材训练生成。具体过程如图2所示：
步骤S201：基于爬虫技术，从目标平台获取目标广告素材。
该目标平台包括搜索平台和/或素材网站。该目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合。
在步骤S201中，基于爬虫技术，从搜索平台上获取所需要进行分析和评价的品牌产品图片，品牌标志logo，品牌广告物料中任一一个进行批量地采集和存储。
例如，要进行评价分析的是A品牌的效果，可以从百度或者谷歌等搜索平台上批量采集A品牌的品牌产品图片，品牌标志logo或品牌广告物料中的任意一个活组合，进行批量存储。
步骤S202：批量标注所述目标广告素材，生成相应的训练集。
在具体实现中，可以利用众包系统或其他的标注工具，如图片标注工具labelimg对所述目标广告素材中包含的各个元素进行批量标注，生成对应各个元素的训练集。也可以利用机器经由学习生成的自动标注工具对目标品牌素材中包含的各个元素进行自动的批量标注，生成对应各个元素的训练集。
例如，将凌渡汽车品牌作为目标广告素材，则需要对与凌渡汽车品牌相关的元素进行标注，如图3所示，通过对图片中的凌渡汽车、凌渡汽车logo，凌渡字样进行标注。
需要说明的是，在本发明中，对于进行批量标注的方式并不仅限于以上公开的两种方式，还可以是其他可进行标注的方式。
步骤S203：对所述训练集进行训练，得到对应的预设模型。
在具体实现中，对上述对应各个元素的训练集，可以基于tensorflow和Faster-RCNN架构对对应各个元素的训练集进行训练，得到对应的预设模型。
其中，tensorflow是一种人工智能学习系统，主要被用于语音识别或图像识别等多项机器深度学习领域。
Faster-RCNN架构则是用于实现目标检测。具体进行目标检测的原理为：先将图像进行归一化处理，从中提取候选区域，然后使用深度网络从提取的候选区域中提取特征，并调整候选框的位置。
其中，在Faster-RCNN架构中的提取特征的操作，可以使用inception、Resnet等模型进行特征提取，使得模型的训练效果加精准。
通过上述方式训练得到的预设模型，可以自适应识别一张图片中的与目标广告素材相关的各个元素。
步骤S103：计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果。
在步骤S103中，有效数据包括：显示时长、次数和面积数据等。
在具体实现中，可以仅计算显示时长，显示次数和面积中的任一一个，或者全部。计算的有效数据越多所得到的计算结果则对后续广告效果的分析越有利。
针对每个元素都计算其在待识别视频中显示时的有效数据。例如，计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
步骤S104：对所述计算结果进行汇总和统计，得到广告效果分析结果。
在具体实现中，通过对得到的计算结果进行汇总和统计，可以形成在待识别视频中与目标广告素材相关的元素的变化情况。如图4所示，为冠名logo在待识别视频中的变化情况。图4中，横坐标为视频时间，纵坐标为冠名logo占比每一帧视频图像整体的面积比率。占比越大代表冠名logo在图像中曝光的越明显。
同样，也可以得到针对其他与目标广告素材相关的元素在待识别视频中的变化情况，
上述步骤S101至步骤S104为对一个视频进行分析的过程，若需要对同一个节目 的所有视频中的广告效果进行分析，则可采用上述步骤S101至步骤S104对所有视频进行分析。
本发明实施例公开的一种广告效果分析方法，通过按帧拆解待识别视频，获取待识别图像集合，然后，基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；再计算确定的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；最后，对所述计算结果进行汇总和统计，得到广告效果分析结果。本发明通过识别待识别视频中与目标广告素材相关的元素，并计算确定该元素在待识别视频中显示时的有效数据，基于该有效数据进行统计和分析，得到体现广告效果的分析结果，实现对现有视频流媒体中的广告效果进行量化评估和分析的目的。进一步，使得企业能够了解到自己的企业品牌在视频中显示时所能够带来的广告效果。
基于上述本发明实施例公开的广告效果分析方法，本发明实施例还对应公开了一种广告效果分析装置，如图5所示，该广告效果分析装置500主要包括：
拆解单元501，用于按帧拆解待识别视频，获取待识别图像集合。
标注单元502，用于基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成。
计算单元503，用于计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果。
可选的，所述有效数据包括：显示时长、次数和面积数据，所述计算单元503，用于计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
分析单元504，用于对所述计算结果进行汇总和统计，得到广告效果分析结果。
进一步的，该广告效果分析装置中还包括：预设单元505。该预设单元505包括：
搜索单元，用于基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站。
批量批注单元，用于批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合。
可选的，批量批注单元可以利用众包系统或labelimg工具对所述目标广告素材中包含的元素进行批量标注，生成对应各个元素的训练集。
可选的，批量批注单元也可以利用机器经由学习生成的自动标注工具对目标品牌素材中包含的各个元素进行自动的批量标注，生成对应各个元素的训练集。
训练单元，用于对所述训练集进行训练，得到对应的预设模型。
可选的，训练单元可以基于tensorflow和Faster-RCNN架构对对应各个元素的所述训练集进行训练，得到对应的预设模型。
上述本发明实施例公开的广告效果分析装置中的各个单元具体的原理和执行过程，与上述本发明实施例公开的广告效果分析方法相同，可参见上述本发明实施例公开的广告效果分析方法中相应的部分，这里不再进行赘述。
基于上述本发明实施例公开的广告效果分析装置，上述各个单元可以通过一种由处理器和存储器构成的硬件设备实现。具体为：上述各个单元和模块作为程序单元存储于存储器中，由处理器执行存储在存储器中的上述程序单元来实现广告效果的分析。
其中，处理器中包含内核，由内核去存储器中调取相应的程序单元。内核可以设置一个或以上，通过调整内核参数来实现对广告效果的分析。
存储器可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)，存储器包括至少一个存储芯片。
进一步的，本发明实施例提供了一种处理器，所述处理器用于运行程序，其中，所述程序运行时执行所述广告效果分析方法。
进一步的，本发明实施例提供了一种设备，设备包括处理器、存储器及存储在存储器上并可在处理器上运行的程序，处理器执行程序时实现以下步骤：按帧拆解待识别视频，获取待识别图像集合；基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；对所述计算结果进行汇总和统计，得到广告效果分析结果。
可选的，所述待识别视频包括离线视频，直播视频和/或动态视频图。
其中，所述有效数据包括：显示时长、次数和面积数据。相应地，计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
其中，所述预设模型由所述目标广告素材训练生成的过程包括：基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；对所述训练集进行训练，得到对应的预设模型。
可选的，可以利用众包系统或labelimg工具对所述目标广告素材中包含的元素进行批量标注，生成对应各个元素的训练集。
可选的，也可以利用机器经由学习生成的自动标注工具对目标品牌素材中包含的各个元素进行自动的批量标注，生成对应各个元素的训练集。
可选的，可以基于tensorflow和Faster-RCNN架构对对应各个元素的所述训练集进行训练，得到对应的预设模型。
本发明实施例中公开的设备可以是服务器、PC、PAD、手机等。
进一步的，本发明实施例还提供了一种存储介质，其上存储有程序，该程序被处理器执行时实现所述广告效果分析方法。
本申请还提供了一种计算机程序产品，当在数据处理设备上执行时，适于执行初始化有如下方法步骤的程序：按帧拆解待识别视频，获取待识别图像集合；基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；对所述计算结果进行汇总和统计，得到广告效果分析结果。
可选的，所述待识别视频包括离线视频，直播视频和/或动态视频图。
其中，所述有效数据包括：显示时长、次数和面积数据。相应地，计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
其中，所述预设模型由所述目标广告素材训练生成的过程包括：基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；对所述训练集进行训练，得到对应的预设模型。
可选的，可以利用众包系统或labelimg工具对所述目标广告素材中包含的元素进 行批量标注，生成对应各个元素的训练集。
可选的，可以基于tensorflow和Faster-RCNN架构对对应各个元素的所述训练集进行训练，得到对应的预设模型。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
存储器可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。存储器是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他 数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白，本申请的实施例可提供为方法、系统或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
以上仅为本申请的实施例而已，并不用于限制本申请。对于本领域技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本申请的权利要求范围之内。

Claims (10)

  1. 一种广告效果分析方法，其特征在于，包括：
    按帧拆解待识别视频，获取待识别图像集合；
    基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；
    计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；
    对所述计算结果进行汇总和统计，得到广告效果分析结果。
  2. 根据权利要求1所述的方法，其特征在于，所述预设模型由所述目标广告素材训练生成，包括：
    基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；
    批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；
    对所述训练集进行训练，得到对应的预设模型。
  3. 根据权利要求2所述的方法，其特征在于，所述对所述训练集进行训练，得到对应的预设模型，包括：
    基于tensorflow和Faster-RCNN架构对对应各个元素的所述训练集进行训练，得到对应的预设模型。
  4. 根据权利要求1-3中任一项所述的方法，其特征在于，所述有效数据包括：显示时长、次数和面积数据，计算确定的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果，包括：
    计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
  5. 根据权利要求1-3中任一项所述的方法，其特征在于，所述待识别视频包括离线视频，直播视频和/或动态视频图。
  6. 一种广告效果分析装置，其特征在于，包括：
    拆解单元，用于按帧拆解待识别视频，获取待识别图像集合；
    标注单元，用于基于预先确定的预设模型对所述待识别图像集合进行识别，确定与目标广告素材相关的元素，并标注，所述预设模型由所述目标广告素材训练生成；
    计算单元，用于计算标注的各个所述元素在所述待识别视频中显示时的有效数据，得到对应的计算结果；
    分析单元，用于对所述计算结果进行汇总和统计，得到广告效果分析结果。
  7. 根据权利要求6所述的装置，其特征在于，所述装置还包括：
    搜索单元，用于基于爬虫技术，从目标平台获取目标广告素材，所述目标平台包括搜索平台和/或素材网站；
    批量批注单元，用于批量标注所述目标广告素材，生成相应的训练集，所述目标广告素材包括品牌产品图片，品牌标志logo，品牌广告物料中的任一一个元素或任意元素的组合；
    训练单元，用于对所述训练集进行训练，得到对应的预设模型。
  8. 根据权利要求6所述的装置，其特征在于，所述有效数据包括：显示时长、次数和面积数据，所述计算单元503，用于计算确定的各个所述元素在所述待识别视频中的显示时长、次数和面积数据，将得到的所述显示时长、所述次数和所述面积数据作为计算结果。
  9. 一种存储介质，其特征在于，所述存储介质包括存储的程序，其中，在所述程序运行时控制所述存储介质所在设备执行如权利要求1-5中任一项所述的广告效果分析方法。
  10. 一种处理器，其特征在于，所述处理器用于运行程序，其中，所述程序运行时执行如权利要求1-5中任一项所述的广告效果分析方法。
PCT/CN2018/101634 2017-09-30 2018-08-22 广告效果分析方法及装置 WO2019062388A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710918855.6 2017-09-30
CN201710918855.6A CN109598527A (zh) 2017-09-30 2017-09-30 广告效果分析方法及装置

Publications (1)

Publication Number Publication Date
WO2019062388A1 true WO2019062388A1 (zh) 2019-04-04

Family

ID=65900482

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/101634 WO2019062388A1 (zh) 2017-09-30 2018-08-22 广告效果分析方法及装置

Country Status (2)

Country Link
CN (1) CN109598527A (zh)
WO (1) WO2019062388A1 (zh)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110348910B (zh) * 2019-07-17 2022-03-01 秒针信息技术有限公司 有效展示时长的确定方法、装置、电子设备及存储介质
CN112446720B (zh) * 2019-08-29 2024-05-14 北京搜狗科技发展有限公司 一种广告显示方法及装置
CN111709762B (zh) * 2020-06-09 2021-10-29 上海极链网络科技有限公司 信息匹配度的评估方法、装置、设备及存储介质
CN111898912A (zh) * 2020-07-31 2020-11-06 成都新潮传媒集团有限公司 一种预选待审广告素材的方法、装置及计算机设备
CN112348566A (zh) * 2020-10-15 2021-02-09 北京捷通华声科技股份有限公司 推荐广告的确定方法、装置及存储介质
CN112598448A (zh) * 2020-12-29 2021-04-02 恩亿科(北京)数据科技有限公司 动态素材数据处理方法、系统、计算机和可读存储介质
CN113112300A (zh) * 2021-04-19 2021-07-13 两比特(北京)科技有限公司 一种影视广告抓取和分析系统
CN113191293B (zh) * 2021-05-11 2023-04-07 创新奇智(重庆)科技有限公司 广告检测方法、装置、电子设备、系统及可读存储介质
CN113743281A (zh) * 2021-08-30 2021-12-03 上海明略人工智能(集团)有限公司 节目广告素材识别方法、系统、计算机设备和存储介质
CN113988949A (zh) * 2021-11-15 2022-01-28 北京有竹居网络技术有限公司 一种推广信息处理方法、装置、设备及介质、程序产品
CN114979691B (zh) * 2022-05-23 2023-07-28 上海影谱科技有限公司 一种体育赛事转播权益广告统计分析方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103518213A (zh) * 2011-03-10 2014-01-15 开放电视公司 广告影响的确定
CN104077577A (zh) * 2014-07-03 2014-10-01 浙江大学 一种基于卷积神经网络的商标检测方法
CN106296292A (zh) * 2016-08-12 2017-01-04 百度在线网络技术（北京）有限公司 用于自动评估投放信息的投放效果的方法和装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2360581A (en) * 2000-03-15 2001-09-26 Television Monitoring Services Quantification of brand exposure on screen
KR100892847B1 (ko) * 2007-05-29 2009-04-10 엔에이치엔(주) 광고 집행에 따른 반응정보 제공 방법 및 시스템
US8340498B1 (en) * 2009-02-03 2012-12-25 Amazon Technologies, Inc. Extraction of text elements from video content
US20110047163A1 (en) * 2009-08-24 2011-02-24 Google Inc. Relevance-Based Image Selection
CN106899864A (zh) * 2015-12-18 2017-06-27 北京国双科技有限公司 广告检测方法和装置
CN107016344A (zh) * 2017-03-08 2017-08-04 上海极链网络科技有限公司 视频中品牌识别系统及其实现方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103518213A (zh) * 2011-03-10 2014-01-15 开放电视公司 广告影响的确定
CN104077577A (zh) * 2014-07-03 2014-10-01 浙江大学 一种基于卷积神经网络的商标检测方法
CN106296292A (zh) * 2016-08-12 2017-01-04 百度在线网络技术（北京）有限公司 用于自动评估投放信息的投放效果的方法和装置

Also Published As

Publication number Publication date
CN109598527A (zh) 2019-04-09

Similar Documents

Publication Publication Date Title
WO2019062388A1 (zh) 广告效果分析方法及装置
CN109117831B (zh) 物体检测网络的训练方法和装置
CN109117848B (zh) 一种文本行字符识别方法、装置、介质和电子设备
TWI753044B (zh) 影片資料的推薦方法和裝置
WO2019223361A1 (zh) 一种视频分析方法及装置
US9785719B2 (en) Generating synthetic data
CN106649316B (zh) 一种视频推送方法及装置
JP2018526710A (ja) 情報推薦方法および情報推薦装置
KR102047568B1 (ko) 웹페이지 렌더링 시간 측정 기법
US20170255830A1 (en) Method, apparatus, and system for identifying objects in video images and displaying information of same
CN107358247B (zh) 一种确定流失用户的方法及装置
EP2899671A1 (en) Identification method and device for target object in image
CN109561052B (zh) 网站异常流量的检测方法及装置
CN108280843A (zh) 一种视频目标检测跟踪方法及设备
CN109255767B (zh) 图像处理方法和装置
US20220189008A1 (en) Method for detecting data defects and computing device utilizing method
US20180075066A1 (en) Method and apparatus for displaying electronic photo, and mobile device
WO2019062841A1 (zh) 品牌曝光效果分析方法及装置
CN109002443A (zh) 一种文本信息的分类方法及装置
CN114730486B (zh) 用于生成用于对象检测的训练数据的方法和系统
CN110569429A (zh) 一种内容选择模型的生成方法、装置和设备
CN108399321B (zh) 基于动态指令依赖图胎记的软件局部抄袭检测方法
US20160188680A1 (en) Electronic device and information searching method for the electronic device
Ragusa et al. Semantic object segmentation in cultural sites using real and synthetic data
Mohsin et al. Convolutional neural networks for real-time wood plank detection and defect segmentation

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: 18860533

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18860533

Country of ref document: EP

Kind code of ref document: A1