SDK overview

更新时间:
复制 MD 格式

Alibaba Cloud Visual Intelligence API offers visual AI technologies across more than 10 categories, such as face and body recognition, OCR, and product understanding. You can use an SDK to call hundreds of features.

Note

If you have questions about SDK integration, API usage, or other issues with the features of Visual Intelligence API, you can contact us by joining our DingTalk support group. The group ID is 23109592.

Prerequisites

SDK languages

The Visual Intelligence API SDK supports six common programming languages: Java, Python, Go, Node.js, PHP, and C#. You can choose the SDK for your language to integrate the visual AI features.

For each category, the following table provides sample code for popular features in common languages and scenarios.

Category

Popular features

Face and body

Features such as 1:N face search, face fusion, and face liveness detection.

OCR

Features such as ID card recognition, driver's license recognition, and PDF recognition.

Product understanding

Product classification.

Content moderation

Text moderation and image moderation.

Image recognition

Features such as general image compliance, element detection, and scene recognition.

Image generation

Features such as image colorization, caption removal, and text-to-image.

Segmentation and matting

Features such as human segmentation, profile picture segmentation, and product segmentation.

Object detection

Features such as entity detection, object detection, and cat and mouse detection.

Video understanding

Shot boundary detection, video OCR, and video content understanding.

Video generation

Features such as general video generation, video face fusion, and video caption removal.

Video segmentation

Green screen video segmentation.

Other scenarios

In addition to server-side integration, Visual Intelligence API also supports the following scenarios:

Note

Place your code on your server to avoid exposing your AccessKey ID and AccessKey secret. These credentials are required to call Visual Intelligence API features. If you make calls from other clients, you must assess the associated security risks.

Calling by signing requests

SDKs are not available for all programming languages. If an SDK is not available for your language, you can call the API by signing your requests. For more information, see Request signing.

Note

When you use the request signing mechanism, you must use an OSS URL in the China (Shanghai) region for the file parameter. For more information about uploading files to OSS, see Activate OSS. If you use a local file or another type of link, you must convert it to an OSS URL in the China (Shanghai) region. For more information, see Method 2 in Handle file URLs.