Overview

更新时间:
复制 MD 格式

OSS Content Moderation detects inappropriate content, such as pornography and politically sensitive material, in images, videos, and audio files stored in Alibaba Cloud Object Storage Service (OSS). This service automatically freezes detected non-compliant content by blocking its public access. This helps you mitigate compliance risks.

Features

OSS Content Moderation scans both new and existing content in a specified OSS bucket. It offers two scan modes: incremental scan and full scan. You can use this feature without writing code by configuring it in the console.
  • Incremental scan

    Once enabled, Content Moderation automatically performs an incremental scan on new files as they are uploaded to an OSS bucket.

  • Full scan

    A full scan is a manually created, one-time task that scans all existing files in a specified bucket at a scheduled time.

Limitations

To use OSS Content Moderation, you must have an active Object Storage Service (OSS) account. For more information about OSS, see What is Object Storage Service?.

Before using OSS Content Moderation, you must authorize it to access your OSS buckets.

Note The first time you sign in to the Content Moderation console and navigate to the OSS Content Moderation page, you can complete this authorization with a single click. For detailed instructions, see Authorize Content Moderation to access OSS buckets.
Table 1. OSS Content Moderation specifications
ItemDescription
Region of the OSS bucketOSS Content Moderation supports OSS buckets in the following regions:
  • oss-cn-hangzhou.aliyuncs.com: The endpoint for OSS buckets in the China (Hangzhou) region.
  • oss-cn-shanghai.aliyuncs.com: The endpoint for OSS buckets in the China (Shanghai) region.
  • oss-cn-qingdao.aliyuncs.com: The endpoint for OSS buckets in the China (Qingdao) region.
  • oss-cn-beijing.aliyuncs.com: The endpoint for OSS buckets in the China (Beijing) region.
  • oss-cn-shenzhen.aliyuncs.com: The endpoint for OSS buckets in the China (Shenzhen) region.
Note For more information about regions, see Regions and endpoints. If your OSS bucket is in another region, you cannot use OSS Content Moderation.
Detection scenarioOSS Content Moderation supports the following detection scenarios:
  • Image moderation scenarios: pornography, terrorism and politically sensitive content, undesirable scenes, and in-image text violations.
  • Video moderation scenarios: pornography, terrorism and politically sensitive content, undesirable scenes, and in-image text violations.
  • Audio moderation scenarios: audio spam.
Supported file typesOSS Content Moderation supports the following image, video, and audio file formats:
  • Image file formats: JPG, PNG, JPEG, GIF, BMP, and WEBP.
  • Video file formats: AVI, MP4, 3GP, MKV, MPG, MPEG, TS, RMVB, WMV, FLV, and MOV.
  • Audio file formats: MP3, WAV, AAC, WMA, OGG, M4A, AMR, AUDIO, and M3U8.

Workflow

StepActionsDescription
Step 1Authorize Content Moderation to access OSS bucketsYou must have an active OSS account and authorize Content Moderation to read your OSS buckets.

For detailed instructions, see Authorize Content Moderation to access OSS buckets.

Note This is a one-time authorization required only when you first use OSS Content Moderation.
Step 2Configure callback notificationsConfigure callback notifications to receive and integrate scan results, which are sent via asynchronous notifications.

For detailed instructions, see Configure notifications.

Step 3Configure risk librariesTo align detection results more closely with your business requirements, Content Moderation lets you configure a custom risk library. You can use this library to specify images that should be blocked, allowed, or flagged for manual review.

For detailed instructions, see Configure risk libraries.

Step 4Configure an incremental scan taskConfigure an incremental scan to automatically check new files as they are added to a specified OSS bucket.

For detailed instructions, see Configure an incremental scan task.

Configure a full scan taskA full scan task performs a one-time scan of all existing files in a specified OSS bucket. After the scan is complete, you can view the results and statistics.

For detailed instructions, see Configure a full scan task.