WO2021258972A1 - 视频检索方法、装置、电子设备和计算机可读介质 - Google Patents

视频检索方法、装置、电子设备和计算机可读介质 Download PDF

Info

Publication number
WO2021258972A1
WO2021258972A1 PCT/CN2021/096148 CN2021096148W WO2021258972A1 WO 2021258972 A1 WO2021258972 A1 WO 2021258972A1 CN 2021096148 W CN2021096148 W CN 2021096148W WO 2021258972 A1 WO2021258972 A1 WO 2021258972A1
Authority
WO
WIPO (PCT)
Prior art keywords
video
retrieval
cluster
videos
search
Prior art date
Application number
PCT/CN2021/096148
Other languages
English (en)
French (fr)
Inventor
孙婉轩
赵嘉伟
王长虎
Original Assignee
北京字节跳动网络技术有限公司
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 北京字节跳动网络技术有限公司 filed Critical 北京字节跳动网络技术有限公司
Publication of WO2021258972A1 publication Critical patent/WO2021258972A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/73—Querying
    • G06F16/732—Query formulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/73—Querying
    • G06F16/735—Filtering based on additional data, e.g. user or group profiles
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the embodiments of the present disclosure relate to the field of computer technology, in particular to video retrieval methods, devices, electronic equipment, and computer-readable media.
  • Video retrieval is common in various scenarios such as video search, video deduplication, and removal of infringing videos.
  • edited videos for example, editing, adding special effects
  • the general retrieval method is easy to omit these videos, resulting in inaccurate retrieval results.
  • Some embodiments of the present disclosure propose video retrieval methods, devices, electronic equipment, and computer-readable media to solve the technical problems mentioned in the background art section above.
  • some embodiments of the present disclosure provide a video retrieval method, including: searching in a video library based on a target video set to obtain a retrieval result; for the video in the retrieval result, generating a related video corresponding to the video to obtain The related video collection corresponding to the retrieval result; searching in the video library based on the related video collection to determine whether there is a video that meets the preset conditions.
  • some embodiments of the present disclosure provide a video retrieval device, including: a first retrieval unit configured to perform retrieval in a video library based on a target video set to obtain retrieval results; and a generating unit configured to The video in the search result is generated, and the related video corresponding to the video is generated, and the related video set corresponding to the search result is obtained; the second retrieval unit is configured to search in the video library based on the related video set to determine whether there is a video that meets the preset conditions .
  • some embodiments of the present disclosure provide an electronic device, including: one or more processors; a storage device, on which one or more programs are stored. When one or more programs are stored by one or more The processor executes, so that one or more processors implement any of the foregoing methods.
  • One of the above-mentioned various embodiments of the present disclosure has the following beneficial effects: for the retrieval result, a corresponding related video is generated, and then based on the related video retrieval, the retrieval accuracy is improved.
  • the associated video can simulate the video obtained after various editing (such as editing, adding special effects) to the video. Therefore, searching based on related videos makes it easier to retrieve edited videos, thereby improving the accuracy of search results.
  • Fig. 1 is a schematic diagram of an application scenario of a video retrieval method according to some embodiments of the present disclosure
  • FIG. 2 is a flowchart of some embodiments of video retrieval methods according to the present disclosure
  • FIG. 3 is a flowchart of other embodiments of the video retrieval method according to the present disclosure.
  • Figure 5 is a schematic structural diagram of some embodiments of a video retrieval device according to the present disclosure.
  • Fig. 6 is a schematic structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure.
  • FIG. 1 is a schematic diagram 100 of an application scenario of a video retrieval method according to some embodiments of the present disclosure.
  • the video retrieval method provided by some embodiments of the present disclosure may be executed by a terminal device, or may be executed by a server.
  • the terminal device and server can be hardware or software.
  • it can be various electronic devices, including but not limited to smart phones, tablet computers, e-book readers, vehicle-mounted terminals, and so on.
  • it can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules for providing distributed services, or as a single software or software module. There is no specific limitation here.
  • the execution subject of the video retrieval method may be the server 101.
  • the server 101 may first perform a search in the video library 103 based on the target video collection 102 to obtain the search result 104.
  • the videos in the target video set 102 may be popular videos within a period of time (for example, within the past three hours).
  • a popular video may be a video with interactive data (for example, the amount of forwarding, the amount of likes, the amount of comments, etc.) greater than a preset threshold.
  • the popular videos are retrieved as the target video collection because the more popular videos are more likely to be infringed.
  • the server 101 can generate an associated video set 105 corresponding to the retrieval result 104.
  • operations such as zooming, cropping, bordering, removing borders, trifolding, etc. may be performed on the video in the search result 104 to obtain the associated video.
  • the server 101 may perform a search in the video library 103 based on the associated video set 105 to determine whether there is a video that meets a preset condition. As an example, it can be determined whether there is a video whose publishing time is later than the video in the target video collection 102.
  • these videos can be determined as infringing videos, and the infringing videos can be deleted.
  • the video retrieval method includes the following steps:
  • Step 201 Search in a video library based on the target video collection to obtain a search result.
  • the execution subject of the video retrieval method may perform retrieval in the video library based on the target video collection, so as to obtain the retrieval result.
  • at least one video in the target video set may be matched in a video library to obtain a retrieval video corresponding to each video in the at least one video.
  • the retrieval video corresponding to each video in at least one video is summarized to obtain the retrieval result.
  • the videos in the target video set can be obtained in various ways. As an example, it can be determined in a specified way, or it can be filtered through certain conditions.
  • the video library can be a collection that includes a large number of videos, which can be specified according to actual needs or filtered according to certain conditions.
  • video matching or retrieval can be performed in various ways. As an example, video matching or retrieval can be performed by calculating the similarity of key frames. As another example, video matching or retrieval can also be performed by calculating the distance between the features of different videos.
  • the above method further includes: based on the interactive data of the video, selecting a video with a preset index greater than a preset threshold from a video library as a target video set.
  • the interactive data of the video includes but is not limited to: the amount of forwarding, the amount of likes, the amount of comments, the amount of playback, and so on.
  • the preset index may be one or more of the interactive data.
  • the target video set is determined based on the interactive data because the higher the interactive data, the more popular these videos can be considered. The more popular videos are edited. In a scene similar to the retrieval of infringing videos, it is easier to retrieve the infringing videos.
  • Step 202 For the video in the search result, generate a related video corresponding to the video, and obtain a related video set corresponding to the search result.
  • the above-mentioned execution subject may generate an associated video corresponding to the video through various methods (for example, data augmentation).
  • multiple associated videos corresponding to these videos can be generated.
  • multiple related videos can be used as a collection of related videos corresponding to the search result.
  • the corresponding associated video can be generated by cropping, zooming, and adding special effects to the video.
  • generating the associated videos corresponding to the videos to obtain the associated video set corresponding to the search results includes: selecting a first preset number of videos from the search results ; For videos in the first preset number of videos, an associated video corresponding to the video is generated, and an associated video set corresponding to the retrieval result is obtained.
  • the first preset number can be any number.
  • the first number of videos can be selected from the search results in various ways. For example, the first preset number of videos may be selected according to the order of the interactive data of the video (for example, the amount of likes, the amount of forwarding, the amount of playback, etc.) in descending order. For another example, the first preset data video can be randomly selected. In these implementations, by selecting a preset number of videos, the workload of generating associated videos can be reduced and the running speed can be improved.
  • Step 203 Search in the video library based on the associated video set, and determine whether there is a video that meets a preset condition.
  • the above-mentioned execution subject may search in the video library based on the associated video set to determine whether there is a video that meets a preset condition. According to actual needs, it can be retrieved in the video library based on one or more videos in the associated video set. In practice, the preset conditions can be different according to actual needs. As an example, for the scene of retrieving infringing videos, it can be determined whether there are videos whose release time is later than the videos in the target video collection. This is because the infringing video is a video obtained by editing on the basis of the original video (video in the target video collection), so its release time is generally later than the original video.
  • the methods provided by some embodiments of the present disclosure generate corresponding related videos for the retrieval results, and then retrieve based on the related videos to improve retrieval accuracy.
  • the associated video can simulate the video obtained after various editing (such as editing, adding special effects) to the video. Therefore, searching based on related videos makes it easier to retrieve edited videos, thereby improving the accuracy of search results.
  • FIG. 3 shows a process 300 of other embodiments of the video retrieval method.
  • the process 300 of the video retrieval method includes the following steps:
  • Step 301 Search in the video library based on the target video collection to obtain the search result.
  • step 301 for the specific implementation of step 301 and the technical effects it brings, reference may be made to step 201 in those embodiments corresponding to FIG. 2, which will not be repeated here.
  • Step 302 Perform visual feature extraction on the retrieval result to obtain a visual feature set.
  • the execution subject of the video retrieval method may perform visual feature extraction on the retrieval result to obtain a visual feature set.
  • visual feature extraction can be performed on part or all of the videos in the retrieval result to obtain a visual feature set.
  • the ORB feature, SIFT feature, etc. of the video can be extracted as visual features.
  • Step 303 Perform clustering on the visual feature set to obtain at least one visual feature cluster.
  • the above-mentioned execution subject may cluster multiple visual features in the visual feature set obtained in step 302 to obtain at least one visual feature cluster.
  • multiple visual features can be clustered through clustering algorithms such as K-means algorithm and mean shift algorithm.
  • clustering algorithms can group similar visual features among multiple visual features into the same cluster, so as to obtain at least one visual feature cluster.
  • Step 304 Obtain at least one video cluster based on the at least one visual feature cluster.
  • the retrieval result can be divided into at least one video cluster according to the corresponding relationship between the visual feature and the video. That is to say, the videos corresponding to each visual feature in the same visual feature cluster are also divided into the same video cluster.
  • Step 305 For video clusters in at least one video cluster, determine the popularity score of the video cluster.
  • the popularity score of the video cluster may be determined.
  • the sum of the popularity scores of each video in the video cluster may be determined as the popularity score of the video cluster.
  • the popularity score of each video can be obtained according to the interactive data (for example, the amount of playback, etc.) of the video. Specifically, as an example, the playback volume can be multiplied by a preset coefficient to obtain the popularity score of the video.
  • the average value of the popularity score of each video in the video cluster may also be determined as the popularity score of the video cluster.
  • Step 306 Select a second preset number of video clusters from at least one video cluster according to the descending order of the popularity score.
  • the above-mentioned execution subject may select a second preset number of video clusters from at least one video cluster according to the order of the popularity score from the largest to the smallest.
  • Step 307 For the videos in the selected video cluster, generate related videos corresponding to the videos, and obtain related video sets corresponding to the retrieval results.
  • the related video corresponding to the video is generated, and the specific implementation of the related video set corresponding to the retrieval result and the technical effects brought by it are obtained.
  • Step 308 Search in the video library based on the associated video set to determine whether there is a video that meets a preset condition.
  • step 308 the specific implementation of step 308 and the technical effects brought by it can refer to step 203 in the embodiment corresponding to FIG. 2, and details are not described herein again.
  • Step 309 In response to determining that there is a video that meets the preset condition, add the video that meets the preset condition to the target video set.
  • the above-mentioned execution subject may add the video that meets the preset condition to the target video set.
  • suitable preset conditions can be set according to actual needs.
  • the preset condition may be: the release time is later than the video in the target video collection.
  • At least one video cluster is obtained by clustering the visual features.
  • the selection is made according to the heat score.
  • the screening of videos in the search results is realized, so as to reduce the number of related videos.
  • the searching speed is improved.
  • the continuous updating of the target video collection is achieved by adding videos that meet the preset conditions into the target video collection. This facilitates subsequent retrieval based on the target video collection, and provides a basis for improving the accuracy of subsequent retrieval.
  • FIG. 4 is a schematic diagram 400 of another application scenario of the video retrieval method according to some embodiments of the present disclosure.
  • the execution subject of the video retrieval method may be the server 401.
  • the server 401 may first search the video library 403 based on the target video collection 402 to obtain the search result 404.
  • visual feature extraction can be performed on the video in the search result 404 to obtain a visual feature set 405.
  • the visual feature set 405 is clustered, and n visual feature clusters 406 are obtained.
  • the retrieval result 404 is also divided into n video clusters 407.
  • the weighted sum of the amount of playback and the number of searches may be determined as the popularity score 408.
  • m video clusters 409 are selected from the n video clusters 407. Generally, m is less than n. Then, for the videos in the m video clusters 409, corresponding related videos are generated, and the related video set 410 is obtained. On this basis, a search is made in the video library 403 based on the associated video set 410 to determine whether there is a video that meets the preset condition. If it exists, videos that meet the preset conditions can be added to the target video set 402.
  • the present disclosure provides some embodiments of a video retrieval device. These device embodiments correspond to those method embodiments shown in FIG. It can be applied to various electronic devices.
  • the video retrieval device 500 of some embodiments includes: a first retrieval unit 501, a generation unit 502, and a second retrieval unit 503.
  • the first retrieval unit 501 is configured to perform retrieval in the video library based on the target video collection to obtain retrieval results.
  • the generating unit 502 is configured to generate a related video corresponding to the video for the video in the search result, and obtain a related video set corresponding to the search result.
  • the second retrieval unit 503 is configured to retrieve in the video library based on the associated video set, and determine whether there is a video that meets a preset condition.
  • the specific implementation of the first retrieval unit 501, the generation unit 502, and the second retrieval unit 503 in the video retrieval device 500 and the technical effects brought about by them can refer to those embodiments corresponding to FIG. 2. Here, No longer.
  • the generating unit 502 may be further configured to select a first preset number of videos from the retrieval result; for videos in the first preset number of videos, generate related videos corresponding to the videos , Get the related video set corresponding to the search result.
  • the generating unit 502 may be further configured to: cluster the retrieval results to obtain at least one video cluster; select a second preset number of video clusters from the at least one video cluster; From the videos in the selected video cluster, the associated videos corresponding to the videos are generated, and the associated video set corresponding to the retrieval result is obtained.
  • the generating unit 502 may be further configured to: perform visual feature extraction on the retrieval result to obtain a visual feature set; cluster the visual feature set to obtain at least one visual feature cluster; For a visual feature cluster, at least one video cluster is obtained.
  • the generating unit 502 may be further configured to: for the video clusters in at least one video cluster, determine the popularity score of the video cluster; according to the order of the popularity score, from at least one Select the second preset number of video clusters from the video clusters.
  • the apparatus 500 may further include: based on the interactive data of the video, selecting a video with a preset index greater than a preset threshold from a video library as the target video set.
  • the apparatus 500 may further include: in response to determining that there is a video that meets the preset condition, adding the video that meets the preset condition into the target video set.
  • a corresponding related video is generated, and then retrieval is based on the related video to improve retrieval accuracy.
  • the associated video can simulate the video obtained after various editing (such as editing, adding special effects) to the video. Therefore, searching based on related videos makes it easier to retrieve edited videos, thereby improving the accuracy of search results.
  • FIG. 6 shows a schematic structural diagram of an electronic device (for example, the server in FIG. 1) 600 suitable for implementing some embodiments of the present disclosure.
  • the electronic device shown in FIG. 6 is only an example, and should not bring any limitation to the function and scope of use of the embodiments of the present disclosure.
  • the electronic device 600 may include a processing device (such as a central processing unit, a graphics processor, etc.) 601, which may be loaded into a random access device according to a program stored in a read-only memory (ROM) 602 or from a storage device 608.
  • the program in the memory (RAM) 603 executes various appropriate actions and processing.
  • various programs and data required for the operation of the electronic device 600 are also stored.
  • the processing device 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604.
  • An input/output (I/O) interface 605 is also connected to the bus 604.
  • the following devices can be connected to the I/O interface 605: including input devices 606 such as touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; including, for example, liquid crystal display (LCD), speakers, vibration An output device 607 such as a device; a storage device 608 such as a magnetic tape, a hard disk, etc.; and a communication device 609.
  • the communication device 609 may allow the electronic device 600 to perform wireless or wired communication with other devices to exchange data.
  • FIG. 6 shows an electronic device 600 having various devices, it should be understood that it is not required to implement or have all of the illustrated devices. It may alternatively be implemented or provided with more or fewer devices. Each block shown in FIG. 6 can represent one device, or can represent multiple devices as needed.
  • the process described above with reference to the flowchart may be implemented as a computer software program.
  • some embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart.
  • the computer program may be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602.
  • the processing device 601 the above-mentioned functions defined in the methods of some embodiments of the present disclosure are executed.
  • the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or a combination of any of the above. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or in combination with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in a baseband or as a part of a carrier wave, and a computer-readable program code is carried therein. This propagated data signal can take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • the computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium.
  • the computer-readable signal medium may send, propagate or transmit the program for use by or in combination with the instruction execution system, apparatus, or device .
  • the program code contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wire, optical cable, RF (Radio Frequency), etc., or any suitable combination of the above.
  • the client and server can communicate with any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and can communicate with digital data in any form or medium.
  • Communication e.g., communication network
  • Examples of communication networks include local area networks (“LAN”), wide area networks (“WAN”), the Internet (for example, the Internet), and end-to-end networks (for example, ad hoc end-to-end networks), as well as any currently known or future research and development network of.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device; or it may exist alone without being assembled into the electronic device.
  • the above-mentioned computer-readable medium carries one or more programs.
  • the electronic device When the above-mentioned one or more programs are executed by the electronic device, the electronic device: performs a search in the video library based on the target video collection to obtain the search result; for the search result In the video, the related video corresponding to the video is generated, and the related video set corresponding to the search result is obtained; based on the related video set, search in the video library to determine whether there is a video that meets the preset conditions.
  • the computer program code used to perform the operations of some embodiments of the present disclosure can be written in one or more programming languages or a combination thereof, the programming languages including object-oriented programming languages-such as Java, Smalltalk, C++ , Also includes conventional procedural programming languages-such as "C" language or similar programming languages.
  • the program code can be executed entirely on the user's computer, partly on the user's computer, executed as an independent software package, partly on the user's computer and partly executed on a remote computer, or entirely executed on the remote computer or server.
  • the remote computer can be connected to the user's computer through any kind of network including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to Connect via the Internet).
  • LAN local area network
  • WAN wide area network
  • each block in the flowchart or block diagram can represent a module, program segment, or part of code, and the module, program segment, or part of code contains one or more for realizing the specified logic function.
  • Executable instructions can also occur in a different order from the order marked in the drawings. For example, two blocks shown one after the other can actually be executed substantially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or operations Or it can be realized by a combination of dedicated hardware and computer instructions.
  • the units described in some embodiments of the present disclosure may be implemented in software or hardware.
  • the described unit can also be provided in the processor, for example, it can be described as: a processor includes a first retrieval unit, a generation unit, and a second retrieval unit. Among them, the names of these units do not constitute a limitation on the unit itself under certain circumstances.
  • the first search unit can also be described as "a unit for searching the target video collection in the video library".
  • exemplary types of hardware logic components include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System on Chip (SOC), Complex Programmable Logical device (CPLD) and so on.
  • FPGA Field Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • ASSP Application Specific Standard Product
  • SOC System on Chip
  • CPLD Complex Programmable Logical device
  • a video retrieval method which includes: performing retrieval in a video library based on a target video set to obtain retrieval results; for the videos in the retrieval results, generating related videos corresponding to the videos, Obtain the associated video set corresponding to the search result; search in the video library based on the associated video set to determine whether there is a video that meets the preset condition.
  • generating the associated video corresponding to the video to obtain the associated video set corresponding to the search result includes: selecting a first preset number of videos from the search result; Among the videos in the first preset number of videos, the associated videos corresponding to the videos are generated, and the associated video set corresponding to the retrieval result is obtained.
  • generating the associated video corresponding to the video to obtain the associated video set corresponding to the search result includes: clustering the search result to obtain at least one video cluster; A second preset number of video clusters are selected from at least one video cluster; for the videos in the selected video cluster, associated videos corresponding to the videos are generated, and an associated video set corresponding to the retrieval result is obtained.
  • clustering the retrieval results to obtain at least one video cluster includes: performing visual feature extraction on the retrieval result to obtain a visual feature set; clustering the visual feature set to obtain at least one Visual feature cluster; based on at least one visual feature cluster, at least one video cluster is obtained.
  • selecting a second preset number of video clusters from at least one video cluster includes: for the video clusters in the at least one video cluster, determining the popularity score of the video cluster; In order of large to small, a second preset number of video clusters are selected from at least one video cluster.
  • the method before searching in the video library based on the target video collection and obtaining the search result, the method further includes: selecting a preset index greater than a preset threshold from the video library based on the interactive data of the video As the target video collection.
  • the method further includes: in response to determining that there is a video that meets the preset condition, adding the video that meets the preset condition to the target video set.
  • a video retrieval device including: a first retrieval unit configured to perform retrieval in a video library based on a target video set to obtain retrieval results; and a generating unit configured to For the videos in the search results, the related videos corresponding to the videos are generated, and the related video sets corresponding to the search results are obtained; the second retrieval unit is configured to search in the video library based on the related video sets to determine whether there are any that meet the preset conditions video.
  • the generating unit may be further configured to select a first preset number of videos from the retrieval result; for videos in the first preset number of videos, generate related videos corresponding to the videos, Get the associated video set corresponding to the retrieval result.
  • the generating unit may be further configured to: cluster the retrieval results to obtain at least one video cluster; select a second preset number of video clusters from the at least one video cluster; The videos in the selected video cluster are generated, and the related videos corresponding to the videos are generated, and the related video sets corresponding to the retrieval results are obtained.
  • the generating unit may be further configured to: perform visual feature extraction on the retrieval result to obtain a visual feature set; cluster the visual feature set to obtain at least one visual feature cluster; Visual feature clusters, at least one video cluster is obtained.
  • the generating unit may be further configured to: for video clusters in at least one video cluster, determine the popularity score of the video cluster; Select the second preset number of video clusters from the clusters.
  • the apparatus may further include: based on the interactive data of the video, selecting a video with a preset index greater than a preset threshold from a video library as a target video set.
  • the apparatus may further include: in response to determining that there is a video that meets the preset condition, adding the video that meets the preset condition to the target video set.
  • an electronic device including: one or more processors; a storage device, on which one or more programs are stored, when one or more programs are Execution by two processors, so that one or more processors implement the method described in any of the foregoing embodiments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种视频检索方法、装置、电子设备和计算机可读介质。该方法包括：基于目标视频集合在视频库中进行检索，得到检索结果（201）；对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合（202）；基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频（203）。该方法实现了提高检索结果的准确性。

Description

视频检索方法、装置、电子设备和计算机可读介质
相关申请的交叉引用
本申请要求于2020年06月23日提交的，申请号为202010579635.7、发明名称为“视频检索方法、装置、电子设备和计算机可读介质”的中国专利申请的优先权，该申请的全文通过引用结合在本申请中。
技术领域
本公开的实施例涉及计算机技术领域，具体涉及视频检索方法、装置、电子设备和计算机可读介质。
背景技术
视频检索常见于视频搜索、视频消重、侵权视频下架等等各种场景。对于一些经过编辑(例如剪辑、添加特效)的视频，一般的检索方法容易将这些视频遗漏，导致检索结果不准确。
发明内容
本公开的内容部分用于以简要的形式介绍构思，这些构思将在后面的具体实施方式部分被详细描述。本公开的内容部分并不旨在标识要求保护的技术方案的关键特征或必要特征，也不旨在用于限制所要求的保护的技术方案的范围。
本公开的一些实施例提出了视频检索方法、装置、电子设备和计算机可读介质，来解决以上背景技术部分提到的技术问题。
第一方面，本公开的一些实施例提供了一种视频检索方法，包括：基于目标视频集合在视频库中进行检索，得到检索结果；对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合；基于关联视频集合在视频库中进行检索，确定是否存在满足 预设条件的视频。
第二方面，本公开的一些实施例提供了一种视频检索装置，包括：第一检索单元，被配置成基于目标视频集合在视频库中进行检索，得到检索结果；生成单元，被配置成对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合；第二检索单元，被配置成基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
第三方面，本公开的一些实施例提供了一种电子设备，包括：一个或多个处理器；存储装置，其上存储有一个或多个程序，当一个或多个程序被一个或多个处理器执行，使得一个或多个处理器实现上述任一的方法。
第四方面，本公开的一些实施例提供了一种计算机可读介质，其上存储有计算机程序，其中，程序被处理器执行时实现上述任一的方法。
本公开的上述各个实施例中的一个实施例具有如下有益效果：对于检索结果，生成对应的关联视频，进而基于关联视频检索，提高检索准确性。其中，由于关联视频可以模拟对视频进行各种编辑(例如剪辑、添加特效)后得到的视频。因此，基于关联视频进行检索，更容易将经过编辑的视频检索出来，从而提高检索结果的准确性。
附图说明
结合附图并参考以下具体实施方式，本公开各实施例的上述和其他特征、优点及方面将变得更加明显。贯穿附图中，相同或相似的附图标记表示相同或相似的元素。应当理解附图是示意性的，原件和元素不一定按照比例绘制。
图1是根据本公开的一些实施例的视频检索方法的一个应用场景的示意图；
图2是根据本公开的视频检索方法的一些实施例的流程图；
图3是根据本公开的视频检索方法的另一些实施例的流程图；
图4是根据本公开的一些实施例的视频检索方法的另一个应用场 景的示意图；
图5是根据本公开的视频检索装置的一些实施例的结构示意图；
图6是适于用来实现本公开的一些实施例的电子设备的结构示意图。
具体实施方式
下面将参照附图更详细地描述本公开的实施例。虽然附图中显示了本公开的某些实施例，然而应当理解的是，本公开可以通过各种形式来实现，而且不应该被解释为限于这里阐述的实施例。相反，提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是，本公开的附图及实施例仅用于示例性作用，并非用于限制本公开的保护范围。
另外还需要说明的是，为了便于描述，附图中仅示出了与有关发明相关的部分。在不冲突的情况下，本公开中的实施例及实施例中的特征可以相互组合。
需要注意，本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分，并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。
需要注意，本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，本领域技术人员应当理解，除非在上下文另有明确指出，否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。
下面将参考附图并结合实施例来详细说明本公开。
图1是根据本公开一些实施例的视频检索方法的一个应用场景的示意图100。
本公开的一些实施例提供的视频检索方法可以由终端设备执行，也可以由服务端执行。需要说明的是，终端设备、服务端可以是硬件，也可以是软件。当其为硬件时，可以是各种电子设备，包括但不限于 智能手机、平板电脑、电子书阅读器、车载终端等等。当其为软件时，可以安装在上述所列举的电子设备中。其可以实现成例如用来提供分布式服务的多个软件或软件模块，也可以实现成单个软件或软件模块。在此不做具体限定。
在本场景下，需要检索视频库中的侵权视频。如图1所示，视频检索方法的执行主体可以是服务器101。服务器101可以首先基于目标视频集合102在视频库103中进行检索，得到检索结果104。作为示例，目标视频集合102中的视频可以是一段时间内(例如过去三小时内)的热门视频。实践中，热门视频可以是交互数据(例如转发量、点赞量、评论量等)大于预设阈值的视频。这里，将热门视频作为目标视频集合进行检索，是因为越热门的视频被侵权的可能性越高。这样，更容易将侵权视频检索出来。在此基础上，服务器101可以生成检索结果104对应的关联视频集合105。作为示例，可以对检索结果104中的视频进行缩放、裁剪、加边框、去边框、三折等等操作，从而得到关联视频。然后，服务器101可以基于关联视频集合105在视频库103中进行检索，确定是否存在满足预设条件的视频。作为示例，可以确定是否存在发布时间晚于目标视频集合102中视频的视频。可选的，若存在，可以将这些视频确定为侵权视频，并将侵权视频进行删除等处理。
继续参考图2，示出了根据本公开的视频检索方法的一些实施例的流程200。该视频检索方法，包括以下步骤：
步骤201，基于目标视频集合在视频库中进行检索，得到检索结果。
在一些实施例中，视频检索方法的执行主体可以基于目标视频集合在视频库中进行检索，从而得到检索结果。具体的，可以将目标视频集合中的至少一个视频在视频库中进行匹配，得到至少一个视频中每个视频对应的检索视频。在此基础上，将至少一个视频中各个视频所对应的检索视频汇总得到检索结果。其中，目标视频集合中的视频可以通过各种方式得到。作为示例，可以通过指定的方式确定，也可 以通过一定的条件筛选得到。视频库可以是包括大量视频的集合，可以根据实际需要进行指定或根据一定的条件进行筛选。
在一些实施例中，可以通过各种方式进行视频匹配或检索。作为示例，可以通过计算关键帧的相似度的方式进行视频匹配或检索。作为又一示例，也可以通过计算不同视频的特征之间的距离的方式进行视频匹配或检索。
在一些实施例的一些可选的实现方式中，在步骤201之前，上述方法还包括：基于视频的交互数据，从视频库中选取预设指标大于预设阈值的视频作为目标视频集合。实践中，视频的交互数据包括但不限于：转发量、点赞量、评论量、播放量等等。预设指标可以是交互数据中的一种或多种。在这些实现方式中，基于交互数据确定目标视频集合，是因为交互数据越高，可以认为这些视频越热门。而热门的视频被编辑的可能性越高。在类似于侵权视频检索的场景下，更容易将侵权视频检索出来。
步骤202，对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。
在一些实施例中，对于检索结果中的某个视频，上述执行主体可以通过各种方法(例如数据扩增)生成该视频对应的关联视频。类似的，对于检索结果中的部分或全部视频，可以生成这些视频对应的多个关联视频。进而，可以将多个关联视频作为检索结果对应的关联视频集合。作为示例，可以通过对视频进行裁剪、缩放、添加特效等处理，生成对应的关联视频。
在一些实施例的一些可选的实现方式中，对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合，包括：从检索结果中选取第一预设数目个视频；对于第一预设数目个视频中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集。在这些实现方式中，第一预设数目可以是任意数目。根据实际需要，可以通过各种方式从检索结果中选取第一数目个视频。例如，可以根据视频的交互数据(例如点赞量、转发量、播放量等)的由大到小的顺序，选取第一预设数目个视频。又如，可以随机选取第一预设数据 个视频。在这些实现方式中，通过选取预设数目个视频，可以减少生成关联视频的工作量，提高运行速度。
步骤203，基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
在一些实施例中，上述执行主体可以基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。根据实际需要，可以基于关联视频集合中的一个或多个视频在视频库中检索。实践中，根据实际需要的不同，预设条件可以有所不同。作为示例，对于检索侵权视频的场景，可以确定是否存在发布时间晚于目标视频集合中视频的视频。这是由于侵权视频是在原视频(目标视频集合中的视频)的基础上进行编辑而得到的视频，因而其发布时间一般晚于原视频。
本公开的一些实施例提供的方法，对于检索结果，生成对应的关联视频，进而基于关联视频检索，提高检索准确性。其中，由于关联视频可以模拟对视频进行各种编辑(例如剪辑、添加特效)后得到的视频。因此，基于关联视频进行检索，更容易将经过编辑的视频检索出来，从而提高检索结果的准确性。
进一步参考图3，其示出了视频检索方法的另一些实施例的流程300。该视频检索方法的流程300，包括以下步骤：
步骤301，基于目标视频集合在视频库中进行检索，得到检索结果。
在一些实施例中，步骤301的具体实现及其所带来的技术效果，可以参考图2对应的那些实施例中的步骤201，在此不再赘述。
步骤302，对检索结果进行视觉特征提取，得到视觉特征集合。
在一些实施例中，视频检索方法的执行主体可以对检索结果进行视觉特征提取，从而得到视觉特征集合。根据实际需要，可以对检索结果中的部分或全部视频进行视觉特征提取，得到视觉特征集合。作为示例，可以提取视频的ORB特征、SIFT特征等作为视觉特征。
步骤303，对视觉特征集合进行聚类，得到至少一个视觉特征簇。
在一些实施例中，上述执行主体可以对步骤302中得到的视觉特 征集合中的多个视觉特征进行聚类，从而得到至少一个视觉特征簇。其中，可以通过K-means算法、均值偏移算法等聚类算法对多个视觉特征进行聚类。各种聚类算法可以将多个视觉特征中类似的视觉特征分到同一个簇，从而得到至少一个视觉特征簇。
步骤304，基于至少一个视觉特征簇，得到至少一个视频簇。
在一些实施例中，由于视觉特征提取自视频，从而可以根据视觉特征与视频的对应关系，将检索结果划分为至少一个视频簇。也就是说，将同一个视觉特征簇中各个视觉特征分别对应的视频也划分到同一个视频簇。
步骤305，对于至少一个视频簇中的视频簇，确定视频簇的热度得分。
在一些实施例中，对于步骤304中得到的至少一个视频簇中的每个视频簇，可以确定该视频簇的热度得分。作为示例，可以将视频簇中各个视频的热度得分之和确定为视频簇的热度得分。其中，每个视频的热度得分可以根据该视频的交互数据(例如播放量等)得到。具体的，作为示例，可以将播放量与预设系数相乘，得到该视频的热度得分。作为又一示例，也可以将视频簇中各个视频的热度得分的平均值确定为视频簇的热度得分。
步骤306，根据热度得分由大到小的顺序，从至少一个视频簇中选取第二预设数量的视频簇。
在一些实施例中，上述执行主体可以根据热度得分由大到小的顺序，从至少一个视频簇中选取第二预设数量的视频簇。
步骤307，对于所选取的视频簇中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。
在一些实施例中，生成视频对应的关联视频，得到检索结果对应的关联视频集合的具体实现及其所带来的技术效果，可以参考图2对应的实施例中的步骤202，在此不再赘述。
步骤308，基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
在一些实施例中，步骤308的具体实现及其所带来的技术效果， 可以参考图2对应的实施例中的步骤203，在此不再赘述。
步骤309，响应于确定存在满足预设条件的视频，将满足预设条件的视频加入目标视频集合。
在一些实施例中，响应于确定存在满足预设条件的视频，上述执行主体可以将满足预设条件的视频加入目标视频集合。其中，可以根据实际需要设定合适的预设条件。作为示例，在检索侵权视频的场景下，预设条件可以是：发布时间晚于目标视频集合中的视频。
从图3中可以看出，与图2对应的一些实施例的描述相比，通过对视觉特征聚类，得到至少一个视频簇。在此基础上，根据热度得分进行选取。从而实现了对于检索结果中的视频的筛选，以减少关联视频的数量。进而在基于关联视频进行检索的过程中，提高了检索速度。此外，通过将满足预设条件的视频加入目标视频集合实现对于目标视频集合的不断更新。从而便于后续基于目标视频集合进行检索，为提高后续检索的准确率提供了基础。
继续参考图4，图4是根据本公开一些实施例的视频检索方法的另一个应用场景的示意图400。
在本应用场景下，视频检索方法的执行主体可以是服务器401。服务器401可以首先基于目标视频集合402在视频库403中进行检索，得到检索结果404。在此基础上，可以对检索结果404中的视频进行视觉特征提取，得到视觉特征集合405。然后，对视觉特征集合405进行聚类，得到n个视觉特征簇406。根据n个视觉特征簇406将检索结果404也划分为n个视频簇407。分别计算每个视频簇的热度得分408。作为示例，可以将播放量与搜量次数的加权和确定为热度得分408。根据热度得分408由大到小的顺序，从n个视频簇407中选取m个视频簇409。一般的，m小于n。然后，对于m个视频簇409中的视频，生成对应的关联视频，得到关联视频集合410。在此基础上，基于关联视频集合410在视频库403中检索，确定否存在满足预设条件的视频。若存在，可以将满足预设条件的视频加入目标视频集合402。
进一步参考图5，作为对上述各图所示方法的实现，本公开提供了一种视频检索装置的一些实施例，这些装置实施例与图2所示的那些方法实施例相对应，该装置具体可以应用于各种电子设备中。
如图5所示，一些实施例的视频检索装置500包括：第一检索单元501、生成单元502和第二检索单元503。其中，第一检索单元501被配置成基于目标视频集合在视频库中进行检索，得到检索结果。生成单元502被配置成对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。第二检索单元503被配置成基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
在一些实施例中，视频检索装置500中的第一检索单元501、生成单元502和第二检索单元503具体实现及其所带来的技术效果，可以参考图2对应的那些实施例，在此不再赘述。
在一些实施例的可选实现方式中，生成单元502可以进一步被配置成从检索结果中选取第一预设数目个视频；对于第一预设数目个视频中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集。
在一些实施例的可选实现方式中，生成单元502可以进一步被配置成：对于检索结果进行聚类，得到至少一个视频簇；从至少一个视频簇中选取第二预设数量的视频簇；对于所选取的视频簇中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。
在一些实施例的可选实现方式中，生成单元502可以进一步被配置成：对检索结果进行视觉特征提取，得到视觉特征集合；对视觉特征集合进行聚类，得到至少一个视觉特征簇；基于至少一个视觉特征簇，得到至少一个视频簇。
在一些实施例的可选实现方式中，生成单元502可以进一步被配置成：对于至少一个视频簇中的视频簇，确定视频簇的热度得分；根据热度得分由大到小的顺序，从至少一个视频簇中选取第二预设数量的视频簇。
在一些实施例的可选实现方式中，装置500还可以包括：基于视频的交互数据，从视频库中选取预设指标大于预设阈值的视频作为目标视频集合。
在一些实施例的可选实现方式中，装置500还可以包括：响应于确定存在满足预设条件的视频，将满足预设条件的视频加入目标视频集合。
在一些实施例中，对于检索结果，生成对应的关联视频，进而基于关联视频检索，提高检索准确性。其中，由于关联视频可以模拟对视频进行各种编辑(例如剪辑、添加特效)后得到的视频。因此，基于关联视频进行检索，更容易将经过编辑的视频检索出来，从而提高检索结果的准确性。
下面参考图6，其示出了适于用来实现本公开的一些实施例的电子设备(例如图1中的服务器)600的结构示意图。图6示出的电子设备仅仅是一个示例，不应对本公开的实施例的功能和使用范围带来任何限制。
如图6所示，电子设备600可以包括处理装置(例如中央处理器、图形处理器等)601，其可以根据存储在只读存储器(ROM)602中的程序或者从存储装置608加载到随机访问存储器(RAM)603中的程序而执行各种适当的动作和处理。在RAM 603中，还存储有电子设备600操作所需的各种程序和数据。处理装置601、ROM 602以及RAM 603通过总线604彼此相连。输入/输出(I/O)接口605也连接至总线604。
通常，以下装置可以连接至I/O接口605：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置606；包括例如液晶显示器(LCD)、扬声器、振动器等的输出装置607；包括例如磁带、硬盘等的存储装置608；以及通信装置609。通信装置609可以允许电子设备600与其他设备进行无线或有线通信以交换数据。虽然图6示出了具有各种装置的电子设备600，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或 更少的装置。图6中示出的每个方框可以代表一个装置，也可以根据需要代表多个装置。
特别地，根据本公开的一些实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的一些实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的一些实施例中，该计算机程序可以通过通信装置609从网络上被下载和安装，或者从存储装置608被安装，或者从ROM 602被安装。在该计算机程序被处理装置601执行时，执行本公开的一些实施例的方法中限定的上述功能。
需要说明的是，本公开的一些实施例所述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的***、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开的一些实施例中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行***、装置或者器件使用或者与其结合使用。而在本公开的一些实施例中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行***、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF(射频)等等，或者上述的任意合适的组合。
在一些实施方式中，客户端、服务器可以利用诸如HTTP(HyperText Transfer Protocol，超文本传输协议)之类的任何当前已知或未来研发的网络协议进行通信，并且可以与任意形式或介质的数字数据通信(例如，通信网络)互连。通信网络的示例包括局域网(“LAN”)，广域网(“WAN”)，网际网(例如，互联网)以及端对端网络(例如，ad hoc端对端网络)，以及任何当前已知或未来研发的网络。
上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备：基于目标视频集合在视频库中进行检索，得到检索结果；对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合；基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的一些实施例的操作的计算机程序代码，所述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(LAN)或广域网(WAN)——连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开各种实施例的***、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。 例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的***来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开的一些实施例中的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的单元也可以设置在处理器中，例如，可以描述为：一种处理器包括第一检索单元、生成单元和第二检索单元。其中，这些单元的名称在某种情况下并不构成对该单元本身的限定，例如，第一检索单元还可以被描述为“于目标视频集合在视频库中进行检索的单元”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、片上***(SOC)、复杂可编程逻辑设备(CPLD)等等。
根据本公开的一个或多个实施例，提供了一种视频检索方法，包括：基于目标视频集合在视频库中进行检索，得到检索结果；对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合；基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
根据本公开的一个或多个实施例，对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合，包括：从检索结果中选取第一预设数目个视频；对于第一预设数目个视频中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集。
根据本公开的一个或多个实施例，对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合，包括：对于检索结果进行聚类，得到至少一个视频簇；从至少一个视频簇中选取第二预设数量的视频簇；对于所选取的视频簇中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。
根据本公开的一个或多个实施例，对于检索结果进行聚类，得到至少一个视频簇，包括：对检索结果进行视觉特征提取，得到视觉特征集合；对视觉特征集合进行聚类，得到至少一个视觉特征簇；基于至少一个视觉特征簇，得到至少一个视频簇。
根据本公开的一个或多个实施例，从至少一个视频簇中选取第二预设数量的视频簇，包括：对于至少一个视频簇中的视频簇，确定视频簇的热度得分；根据热度得分由大到小的顺序，从至少一个视频簇中选取第二预设数量的视频簇。
根据本公开的一个或多个实施例，在基于目标视频集合在视频库中进行检索，得到检索结果之前，方法还包括：基于视频的交互数据，从视频库中选取预设指标大于预设阈值的视频作为目标视频集合。
根据本公开的一个或多个实施例，方法还包括：响应于确定存在满足预设条件的视频，将满足预设条件的视频加入目标视频集合。
根据本公开的一个或多个实施例，提供了一种视频检索装置，包括：第一检索单元，被配置成基于目标视频集合在视频库中进行检索，得到检索结果；生成单元，被配置成对于检索结果中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合；第二检索单元，被配置成基于关联视频集合在视频库中进行检索，确定是否存在满足预设条件的视频。
根据本公开的一个或多个实施例，生成单元可以进一步被配置成从检索结果中选取第一预设数目个视频；对于第一预设数目个视频中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集。
根据本公开的一个或多个实施例，生成单元可以进一步被配置成：对于检索结果进行聚类，得到至少一个视频簇；从至少一个视频簇中选取第二预设数量的视频簇；对于所选取的视频簇中的视频，生成视频对应的关联视频，得到检索结果对应的关联视频集合。
根据本公开的一个或多个实施例，生成单元可以进一步被配置成：对检索结果进行视觉特征提取，得到视觉特征集合；对视觉特征集合进行聚类，得到至少一个视觉特征簇；基于至少一个视觉特征簇，得到至少一个视频簇。
根据本公开的一个或多个实施例，生成单元可以进一步被配置成：对于至少一个视频簇中的视频簇，确定视频簇的热度得分；根据热度得分由大到小的顺序，从至少一个视频簇中选取第二预设数量的视频簇。
根据本公开的一个或多个实施例，装置还可以包括：基于视频的交互数据，从视频库中选取预设指标大于预设阈值的视频作为目标视频集合。
根据本公开的一个或多个实施例，装置还可以包括：响应于确定存在满足预设条件的视频，将满足预设条件的视频加入目标视频集合。
根据本公开的一个或多个实施例，提供了一种电子设备，包括：一个或多个处理器；存储装置，其上存储有一个或多个程序，当一个或多个程序被一个或多个处理器执行，使得一个或多个处理器实现如上述任一实施例描述的方法。
根据本公开的一个或多个实施例，提供了一种计算机可读介质，其上存储有计算机程序，其中，程序被处理器执行时实现如上述任一实施例描述的方法。
以上描述仅为本公开的一些较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开的实施例中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开的实施例中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。

Claims (10)

  1. 一种视频检索方法，包括：
    基于目标视频集合在视频库中进行检索，得到检索结果；
    对于所述检索结果中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集合；
    基于所述关联视频集合在所述视频库中进行检索，确定是否存在满足预设条件的视频。
  2. 根据权利要求1所述的方法，其中，所述对于所述检索结果中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集合，包括：
    从所述检索结果中选取第一预设数目个视频；
    对于所述第一预设数目个视频中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集。
  3. 根据权利要求1所述的方法，其中，所述对于所述检索结果中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集合，包括：
    对于所述检索结果进行聚类，得到至少一个视频簇；
    从所述至少一个视频簇中选取第二预设数量的视频簇；
    对于所选取的视频簇中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集合。
  4. 根据权利要求3所述的方法，其中，所述对于所述检索结果进行聚类，得到至少一个视频簇，包括：
    对所述检索结果进行视觉特征提取，得到视觉特征集合；
    对所述视觉特征集合进行聚类，得到至少一个视觉特征簇；
    基于所述至少一个视觉特征簇，得到所述至少一个视频簇。
  5. 根据权利要求4所述的方法，其中，所述从所述至少一个视频簇中选取第二预设数量的视频簇，包括：
    对于所述至少一个视频簇中的视频簇，确定所述视频簇的热度得分；
    根据热度得分由大到小的顺序，从所述至少一个视频簇中选取所述第二预设数量的视频簇。
  6. 根据权利要求1所述的方法，其中，在所述基于目标视频集合在视频库中进行检索，得到检索结果之前，所述方法还包括：
    基于视频的交互数据，从所述视频库中选取预设指标大于预设阈值的视频作为所述目标视频集合。
  7. 根据权利要求1-5中任一所述的方法，其中，所述方法还包括：
    响应于确定存在满足预设条件的视频，将所述满足预设条件的视频加入所述目标视频集合。
  8. 一种视频检索装置，包括：
    第一检索单元，被配置成基于目标视频集合在视频库中进行检索，得到检索结果；
    生成单元，被配置成对于所述检索结果中的视频，生成所述视频对应的关联视频，得到所述检索结果对应的关联视频集合；
    第二检索单元，被配置成基于所述关联视频集合在所述视频库中进行检索，确定是否存在满足预设条件的视频。
  9. 一种电子设备，包括：
    一个或多个处理器；
    存储装置，其上存储有一个或多个程序，
    当所述一个或多个程序被所述一个或多个处理器执行，使得所述一个或多个处理器实现如权利要求1-7中任一所述的方法。
  10. 一种计算机可读介质，其上存储有计算机程序，其中，所述程序被处理器执行时实现如权利要求1-7中任一所述的方法。
PCT/CN2021/096148 2020-06-23 2021-05-26 视频检索方法、装置、电子设备和计算机可读介质 WO2021258972A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010579635.7A CN111666449B (zh) 2020-06-23 2020-06-23 视频检索方法、装置、电子设备和计算机可读介质
CN202010579635.7 2020-06-23

Publications (1)

Publication Number Publication Date
WO2021258972A1 true WO2021258972A1 (zh) 2021-12-30

Family

ID=72389431

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/096148 WO2021258972A1 (zh) 2020-06-23 2021-05-26 视频检索方法、装置、电子设备和计算机可读介质

Country Status (2)

Country Link
CN (1) CN111666449B (zh)
WO (1) WO2021258972A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111666449B (zh) * 2020-06-23 2023-04-25 抖音视界有限公司 视频检索方法、装置、电子设备和计算机可读介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101071439A (zh) * 2007-05-24 2007-11-14 北京交通大学 一种基于多视角的交互式视频搜索方法
US20140223480A1 (en) * 2012-02-02 2014-08-07 Digitalsmiths, Inc. Ranking User Search and Recommendation Results for Multimedia Assets Using Metadata Analysis
CN108255970A (zh) * 2017-12-26 2018-07-06 努比亚技术有限公司 一种视频查找方法、终端及计算机可读存储介质
CN111666449A (zh) * 2020-06-23 2020-09-15 北京字节跳动网络技术有限公司 视频检索方法、装置、电子设备和计算机可读介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0015896D0 (en) * 2000-06-28 2000-08-23 Twi Interactive Inc Multimedia publishing system
US9152708B1 (en) * 2009-12-14 2015-10-06 Google Inc. Target-video specific co-watched video clusters
CN110929058B (zh) * 2018-08-30 2023-01-31 北京蓝灯鱼智能科技有限公司 商标图片的检索方法、装置、存储介质及电子装置
CN110121079A (zh) * 2019-05-13 2019-08-13 北京百度网讯科技有限公司 视频处理方法、装置、计算机设备及存储介质
CN111241345A (zh) * 2020-02-18 2020-06-05 腾讯科技（深圳）有限公司 一种视频检索方法、装置、电子设备和存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101071439A (zh) * 2007-05-24 2007-11-14 北京交通大学 一种基于多视角的交互式视频搜索方法
US20140223480A1 (en) * 2012-02-02 2014-08-07 Digitalsmiths, Inc. Ranking User Search and Recommendation Results for Multimedia Assets Using Metadata Analysis
CN108255970A (zh) * 2017-12-26 2018-07-06 努比亚技术有限公司 一种视频查找方法、终端及计算机可读存储介质
CN111666449A (zh) * 2020-06-23 2020-09-15 北京字节跳动网络技术有限公司 视频检索方法、装置、电子设备和计算机可读介质

Also Published As

Publication number Publication date
CN111666449A (zh) 2020-09-15
CN111666449B (zh) 2023-04-25

Similar Documents

Publication Publication Date Title
TWI737006B (zh) 一種跨模態訊息檢索方法、裝置和儲存介質
WO2021196903A1 (zh) 视频处理方法、装置、可读介质及电子设备
WO2020156389A1 (zh) 信息推送方法和装置
JP7222008B2 (ja) 動画クリップの検索方法及び装置
WO2020107625A1 (zh) 视频分类方法、装置、电子设备及计算机可读存储介质
WO2022037419A1 (zh) 音频内容识别方法、装置、设备和计算机可读介质
CN109862100B (zh) 用于推送信息的方法和装置
WO2023016349A1 (zh) 一种文本输入方法、装置、电子设备和存储介质
WO2023151589A1 (zh) 视频显示方法、装置、电子设备和存储介质
WO2023142913A1 (zh) 视频处理方法、装置、可读介质及电子设备
WO2020199659A1 (zh) 用于确定推送优先级信息的方法和装置
CN110198473B (zh) 视频处理方法、装置、电子设备及计算机可读存储介质
US20230315990A1 (en) Text detection method and apparatus, electronic device, and storage medium
WO2021258972A1 (zh) 视频检索方法、装置、电子设备和计算机可读介质
WO2021227953A1 (zh) 图像特效配置方法、图像识别方法、装置及电子设备
CN113919320A (zh) 异构图神经网络的早期谣言检测方法、***及设备
WO2020052059A1 (zh) 用于生成信息的方法和装置
US11437038B2 (en) Recognition and restructuring of previously presented materials
JP7288062B2 (ja) 情報を出力するための方法及び装置、電子機器、記憶媒体並びにコンピュータプログラム
WO2024001548A1 (zh) 歌单生成方法、装置、电子设备及存储介质
WO2023000782A1 (zh) 获取视频热点的方法、装置、可读介质和电子设备
CN115631514B (zh) 基于掌静脉指纹的用户识别方法、装置、设备与介质
CN111460211A (zh) 音频信息播放方法、装置和电子设备
CN116958852A (zh) 视频与文本的匹配方法、装置、电子设备和存储介质
WO2022148239A1 (zh) 信息输出方法、装置和电子设备

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: 21829360

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 30/03/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21829360

Country of ref document: EP

Kind code of ref document: A1