object-scene-recognition

To auto-tag images for smart albums, this Skill identifies objects and scenes in photos. It returns up to five ranked labels with confidence scores, eliminating manual inspection and enabling consistent photo categorization.

Object detection via the Baidu API supports Base64 encoded image strings or public image URLs as input. You provide one of these formats, and the API analyzes the photo to return detected scene and object labels.

Yes, you can get additional context for detected image labels. The Skill supports optional Baidu Baike metadata retrieval, providing richer interpretation and descriptive information alongside the standard confidence scores for recognized objects.

Image content moderation works seamlessly with MiniProgram workflows. The Skill supports web and MiniProgram environments, allowing product teams to upload user photos and automatically review or tag visual content.

Yes, you need platform-managed access to the Baidu advanced_general image classification API to perform scene classification. This access is required for the Skill to process your images and return object and scene labels.

A limitation of automated photo tagging for content review is that the Skill returns a maximum of five labels per image. While confidence scores help gauge accuracy, complex images may require additional manual verification for thorough moderation.