Smart tag
Smart Tag analyzes information in videos, such as visual data, text, speech, and behavior. It uses multimodal information fusion and alignment technology to accurately identify content. This feature automatically generates multi-dimensional content tags for videos and converts unstructured information into a structured format. You can use Smart Tag for intelligent video analysis, video moderation, video search, and personalized recommendations to support intelligent video production. This topic describes how to generate smart tags for videos.
Features
Multi-dimensional information identification
Smart tag analyzes multi-dimensional information such as visual information, text, voice, and behavior to identify people, object scenarios, landmarks, text, and voice in videos.
Multimodal understanding
ApsaraVideo VOD uses technologies such as multimodal fusion and alignment and cross-mode mapping to identify media content to efficiently and accurately understand video content.
Abundant and flexible tags
Smart Tag uses a mature media tagging system to add smart tags to media files based on multimodal analysis. This process generates multi-dimensional video tags and offers flexible tag customization.
Scenarios
Accurate and efficient media resource retrieval
Using the accurate and rich video tag system and cataloging media files such as videos, images, and text files, smart tag supports quick retrieval for content in the media library based on the keyword and tag text. This makes the video retrieval more efficient and accurate.
Personalized recommendation
Precise and personalized content recommendation. Smart tag precisely matches content to achieve personalized video recommendation by combining user information, behavior data, and user profile analysis based on the tags of media content files. This solves the cold start problem in the recommendation system and makes the recommendation more accurate.
Smart advertising
Intelligent and scenario-based advertising. Smart tag automatically identifies the advertising space and matches the advertising space with the content of the advertising resource library based on multimodal content analysis of media files and video tags. This allows the ads to be intelligently delivered and audience-targeted, which makes the ad conversion rate higher.
Fees
Smart Tag is billed based on the duration of successfully processed videos. For more information, see Smart Tag billing.
Limits
The Smart Tag feature is available only in the China North 2 (Beijing) and China East 2 (Shanghai) regions.
Prerequisites
Smart tags can be applied only to videos that are uploaded to ApsaraVideo VOD. For more information about how to upload videos, see Media upload.
How to use
Submit a Smart Tag job
After you upload a video to ApsaraVideo VOD, call the Submit a Smart Tag or Media Fingerprint Job operation to submit a Smart Tag job.
Receive callback information
After the Smart Tag job is complete, if the AI Processing Completed event notification is enabled, a message is sent to the callback URL. For more information about the callback message format, see Smart Tag job completed.
Query the result of a job
You can use API operations or SDKs to query job results. You can use a job ID or video ID to query the job results.
To query job information and results by job ID, see Query a Smart Tag or Media Fingerprint Job.
To query results by video ID, see Get Smart Tag results.
SDK examples
For SDK examples for Smart Tag, see Smart Tag management.