A media fingerprint is a binary string that uniquely identifies a video. You can use the media fingerprinting feature to extract and compare fingerprint features, such as images and audio, from videos. This helps you identify duplicate videos and trace the source of video clips. This feature is useful for scenarios such as original content identification, video deduplication, quick review, ad revenue sharing, and video source tracing.
Overview
Scenarios
Original content identification: Accurately identifies whether a video is original, edited, or re-created by independent creators. This helps resolve copyright infringement issues in the short video industry, provides a technical foundation for video material trading ecosystems, and helps prevent infringing videos from harming the interests of original creators and video platforms.
Video deduplication: You can establish an independent media fingerprint library to quickly compare sample videos and recall duplicate videos or clips in real time. This prevents an excessive number of identical or similar videos from negatively affecting the user experience during content delivery and personalized recommendation.
Quick review: You can maintain a library of non-compliant videos or a custom blacklist to quickly compare the fingerprints of incremental videos and determine whether they contain non-compliant content. Compared with traditional review methods, this method improves the efficiency and accuracy of reviewing large volumes of videos and reduces review costs.
Ad revenue sharing: You can use media fingerprinting technology to retrieve ads from a library for revenue sharing, and to monitor and identify specific ads. This supports the dynamic ad revenue sharing ecosystem, allowing you to easily control the time and frequency of ad delivery and protecting the interests of both copyright owners and ad platforms.
NoteThe media fingerprinting feature of ApsaraVideo Media Processing supports the identification of only ads that are already in the media fingerprint library. It cannot identify ads that have not been added to the library. To identify ads, you must first upload the ad materials to the media fingerprint library.
Features
Features: Media fingerprinting
Limits
The media fingerprinting feature is available in the China (Beijing), China (Hangzhou), China (Shanghai), and Singapore regions.
For more information about the concurrency limits on MPS queues, see Limits on MPS queues.
Before you begin
Before you begin, complete the required settings to ensure that the feature works as expected. If you have already completed these settings, you can skip this section.
Create an MPS queue. If a media fingerprinting queue is not available in the queue list, you must create one.
If required, call the CreateFpShotDB operation to create a media fingerprint library.
Upload the video to be processed to OSS.
If you want to create a job, directly upload the video to OSS. For more information, see Upload a video.
If you want to create a workflow and automatically trigger the workflow, you must add media buckets and create a workflow before you upload the video. After you specify the input and output media buckets and upload a media file to the input media bucket, MPS receives a message about the upload and triggers the associated workflow. After the workflow is complete, the processed media file is uploaded to the output media bucket. For more information, see Add media buckets.
Optional. Enable the notification feature. If you want MPS to send notifications on jobs and workflows to MNS topics or queues, configure an MNS queue or topic as required. For more information, see Enable the notification feature for MPS jobs or workflows.
Submit a media fingerprinting job in the console
You cannot specify a media fingerprint library when you create a media fingerprinting job in the ApsaraVideo Media Processing console. By default, the most recently created media fingerprint library is used.
Submit a single media fingerprinting job
Log on to the MPS console.
In the top navigation bar, select a region from the drop-down list.
In the navigation pane on the left, choose Task Management.
Click the Media Fingerprinting tab, and then click Create Job.
Configure the parameters.
Parameter | Description |
Input File URL | Click Select. In the Bucket drop-down list, select a bucket name. The folders in the bucket are displayed under Path. Select a file in a folder as the input path. |
Media Fingerprinting Queue | Select a queue from the drop-down list. Note If no MPS queue name is displayed in the Video DNA Pipeline drop-down list, create an MPS queue for media fingerprinting. For more information, see Create an MPS queue. |
Library Import Rule | Select a rule from the Ingestion Rule options. The default value is Save Non-duplicated Content Only. |
Click Submit Job to create and submit the media fingerprinting job.
After a media fingerprinting job is submitted, you can view the processing result and details on the Tasks > Video DNA tab.
Submit a media fingerprinting job using a workflow
Log on to the MPS console.
In the top navigation bar, select a region from the drop-down list.
In the navigation pane on the left, choose .
Click Create Workflow.
Set basic workflow configurations.
Parameter
Description
Workflow Name
Enter a workflow name in the text box.
Configuration Plan
Select Custom from the drop-down list.
Set trigger and other configurations.
Parameter
Description
Input Path
Click Select and select a bucket from the Bucket drop-down list. The folders in the selected bucket are automatically displayed under Path. Select a folder as the input path.
NoteIf no bucket is available in the drop-down list, add a bucket first. For more information, see Add media buckets.
Transcoding Queue
Select a queue from the drop-down list.
NoteThis queue is used for transcoding and snapshots. The media fingerprinting queue required for the media fingerprinting job is set in a subsequent node.
If no queue is available in the drop-down list, create a queue first. For more information, see Create an MPS queue.
Media Publishing
Set Media Publishing to Automatic Publishing.
Manual Publish: After a workflow is successfully executed, the media is set to an unpublished status. You can then manually publish the media. To manually publish media files, see Manage Media.
Auto-publish: After the workflow is successfully executed, the media is set to the published status.
NotePublished: The access permissions on all playback resources and snapshot files of the media asset inherit the access permissions of the bucket where they are stored.
Unpublished: The access permissions on all playback resources and snapshot files of the media asset are set to private.
(Optional) Notifications
Click the Switch button, select Message Queue or Message Topic, and select a queue notification or topic notification from the drop-down list.
NoteIf no queue or topic for notifications is available in the drop-down list, enable notifications first. For more information, see Configure notifications.
Add a Media Fingerprinting node.
Click the plus icon to the right of the Input node and select Video DNA.
Click the Pen icon to the right of the Video DNA node.
In the Video DNA dialog box, configure the parameters that are described in the following table.
Parameter
Description
Media Fingerprinting Queue
Select a queue from the drop-down list.
NoteIf no MPS queue name is displayed in the Video DNA Pipeline drop-down list, enable an MPS queue for media fingerprinting. For more information, see Create an MPS queue.
Ingestion Rules
Save Non-duplicated Content Only: Adds only the fingerprints of non-duplicate videos to the media fingerprint library.
Do Not Save Any Media Content: Videos are only compared, and their media fingerprints are not added to the library.
Terminate Workflow
Select an option as needed.
Click OK to complete the configuration of the Media Fingerprinting node.
Click Save to create the workflow.
After the workflow is created, it is automatically triggered when a new file that meets the specified conditions is uploaded to the specified path.
After a workflow instance is executed, to view the details of the Video DNA node, click Details in the Actions column for the instance. On the instance details page, find and copy the Task Name. Then, in the navigation pane on the left, go to Task Management > Video DNA and select Search By ID to search for the task.
Submit a media fingerprinting job using an API
Create a media fingerprint library.
Submit a job to create a media fingerprint library. Information about the new library is returned in the response. After the library is created, its status changes to active. For more information about the parameters, see CreateFpShotDB.
NoteA media fingerprint library is used to record and store media fingerprints. The media fingerprinting service supports the following operations:
Create media fingerprint libraries based on user IDs (UIDs).
A single user (identified by a UID) can create multiple media fingerprint libraries for different business scenarios.
Submit a media fingerprinting job.
If you have less than 2,000 hours of existing videos, you can directly submit a job by calling the Submit DNA Job operation. For more information, see Submit DNA Job. If you have a large video library, submit a ticket to request the required resources.
To process new videos in real time, you can also call the SubmitFpShotJob - Submit DNA Job operation to process them. You can use this operation to submit a job that processes both incremental and existing videos. Alternatively, you can wait until the existing videos are imported and then call the Submit DNA Job operation to submit a job.
Query the job results.
Query the job results: You can call the QueryFpShotJobList - Query DNA Job Results operation or use a callback. The results contain the following key information:
Whether duplicate videos are found.
Information about duplicate videos if any are found.
The time code intervals of the duplicate segments between videos.
The degree of duplication between videos. The value is a number between 0 and 1.
Receive callback messages: After the job is complete, if MNS notifications are configured for the MPS queue, a message is sent to the specified Simple Message Queue (formerly MNS) queue or topic. For more information, see Receive notifications.