Console for Video Moderation Enhanced Edition and Video Moderation 2.0

更新时间:
复制 MD 格式

The Content Safety Video Moderation Enhanced Edition service includes default moderation policies based on extensive experience and industry standards. If these defaults do not meet your business needs, you can customize them.

Use cases

If your business has the following requirements, you can use the Video Moderation Enhanced Edition console to customize detection rules and scope, query detection results, and monitor usage statistics.

When using the service for the first time, log on to the Content Moderation console to review the initial threat detection policy configuration.

Use case

Description

Adjust threat detection policies for video moderation

You can adjust video detection policies based on industry practices or your business needs. You can adjust the following policies:

  • Image detection scope

  • Voice detection scope

  • Result return method

For example, to detect only high-risk violations and capture one snapshot every two seconds, you can adjust the detection rules in the console.

Set different threat detection policies for multiple business lines

If your business lines use the same service but require different threat detection policies, you can copy the service to create separate configurations.

For example, if business lines A, B, and C all use the videoDetection (video file moderation) service but have different standards, you can copy the videoDetection service to create videoDetection_01 and videoDetection_02. By setting different threat detection policies for these three services, you can meet the needs of each business line.

Store video snapshots and audio clips

For long-term storage of video snapshots, you can configure the service to save them to your OSS bucket.

Note

By default, the system temporarily stores video snapshots for 30 minutes and returns temporary URLs through the API. The evidence storage feature is available only in regions within the Chinese mainland.

Query detailed video moderation results

To review the details of recently moderated videos, use the Detection Results page to view or search for the results.

Monitor recent video moderation usage

If you need to monitor and view recent video moderation usage, you can do so on the Dashboards page.

Activate the service

Go to the Video Moderation Enhanced Edition page to activate the Video Moderation Enhanced Edition service.

Note

Before you activate the Video Moderation Enhanced Edition service, review the billing rules for Video Moderation Enhanced Edition. For more information, see Introduction and billing of Video Moderation Enhanced Edition V2.0.

Adjust threat detection policies

Adjust the detection scope and result return method based on industry practices or your specific business needs.

  1. Log on to the Content Moderation console.

  2. In the left-side navigation pane, choose Machine Moderation Enhanced Edition > Video Moderation > Rules.

  3. On the rules page, find the video file moderation (videoDetection) service and click Settings in the Actions column.

  4. On the rule configuration page for the video file moderation (videoDetection) service, adjust the configurations as needed.

    Example: To take 1,000 snapshots per video, detect high-risk frame violations and audio violations, and return all detection results, configure the settings as follows:

    1. For the snapshot method, select Fixed total number of snapshots, enter 1000 in the total snapshots box, and select General baseline detection for the image moderation rule.

      For more information about image moderation rules, see Introduction and billing of Image Moderation 2.0.

      image.png

    2. Turn on the Moderate Audio switch and select Audio and video media moderation.

      image.png

    3. Select Return All Results.

      image.png

    4. Click Save to save the new detection rules.

      The new rules take effect in about 2 to 5 minutes.

Set different policies for business lines

Copy a service and configure different threat detection policies for each copy to meet the different moderation requirements of your business lines.

  1. Log on to the Content Moderation console.

  2. In the left-side navigation pane, choose Machine Moderation Enhanced Edition > Video Moderation > Rules.

  3. On the rules page, copy the service for which you want to set different detection scopes.

    1. In the service list, find the video file moderation service (videoDetection) and click Copy in the Actions column.

    2. In the Copy Service dialog box, enter a Service Name and Service Description.

      image.png

    3. Click OK. After the service is created, find the new service (videoDetection_01), click Settings in the Actions column, and adjust the rules as needed.

      The new service (videoDetection_01) is ready to be called 1 to 2 minutes after creation. You can copy multiple services and adjust the rules for each to meet your different threat detection requirements.

Store video snapshots and audio clips

For long-term storage, configure the service to save video snapshots and audio clips to your OSS bucket.

  1. Log on to the Content Moderation console.

  2. In the left-side navigation pane, choose Machine Moderation Enhanced Edition > Video Moderation > Rules.

  3. On the rules page, find the service you want to configure, and click Settings in the Actions column. This example uses the video file moderation (videoDetection) service.

  4. On the rule configuration page for the video file moderation (videoDetection) service, select the Result Configuration tab.

    Example: To store video snapshots and audio clips in the video-001 folder of an OSS bucket and store all results, configure the settings as follows:

    1. Turn on the Store Evidence switch (OSS authorization is required on first use), select the Storage Bucket, select the Content to Store, enter the Storage Directory, set the Access Link Validity Period, and select the Storage Scope.

    2. Click Save to save the evidence storage configuration.

      The new rule takes effect in about 2 to 5 minutes.

Query detailed video moderation results

You can query moderation results for specific videos by using filters such as taskId, dataId, service, image label, voice label, or threat level.

  1. Log on to the Content Moderation console.

  2. In the left-side navigation pane, choose Machine Moderation Enhanced Edition > Video Moderation > Detection Results.

  3. On the Detection Results page, enter query conditions to search for detection results.

    To search for all records with a matching image label, set the query condition to Image Label is not equal to nonLabel. You can specify multiple labels separated by commas (,).

    Note

    By default, results are displayed in reverse chronological order, with a maximum of 50,000 records shown. The Content Moderation console retains query results for the last 30 days. For long-term analysis, we recommend storing the data or logs from each API call.

    • Click Details in the Actions column of a result to view its moderation information.

    • Click Audio and Image Results in the Actions column to view the detailed moderation results for video frames and audio.

      The Image Moderation Results tab displays each snapshot and its matched labels.

      image

      The Voice Moderation Results tab displays the speech-to-text transcript and matched labels.

      image

      Note

      Voice moderation results are displayed only if you enabled voice moderation in the video moderation settings.

Monitor usage statistics

Monitoring recent video moderation usage helps you develop further review or governance strategies for specific video content.

  1. Log on to the Content Moderation console.

  2. In the left-side navigation pane, choose Machine Moderation Enhanced Edition > Video Moderation > Dashboards.

  3. On the Dashboards page, select a time range to query or export usage data.

    • Query usage: You can view usage statistics by day or month. Data is stored for one year, and you can query a period of up to two months at a time. You can filter the data by Alibaba Cloud account, RAM user, or service.

      • The line chart shows the number of video moderation tasks (number of videos).

      • The bar chart shows the number of calls for image and voice services, which includes the number of moderated video snapshots and the total audio moderation duration in minutes.

    • Export usage: Click the image.png icon to export usage data by day or month. The exported report is in Excel format and includes data only for periods with call volume. The report fields are described below:

      Parameter

      Description

      Unit

      Account UID

      The UID of the account for which data is exported.

      N/A

      service

      • Video moderation service

      • Video image moderation service

      • Video voice moderation service

      N/A

      Number of calls

      The total number of calls.

      N/A

      Date

      The date of the statistics.

      Day/Month

    • View video moderation threat statistics: This section displays statistics on threat levels (high, medium, low, and none), as well as threat trends and distribution for the selected time range.

      • Threat Level Distribution: Shows the count and percentage of high-threat, medium-threat, low-threat, and no-threat results for the selected time range.

      • Threat Trend Statistics: Shows the trend of threats by day or month for the selected time range.

      • Threat Distribution Statistics: Shows threat statistics for video frames and audio. You can view the top 5 threats and the proportion of all threats.