Help with Object Detection and Bounding Box Extraction Using Azure Content Understanding

Hello,

I’m currently working on a video analysis project using Azure Content Understanding. My goal is to detect objects in video files and extract their bounding box positions (coordinates within the frame).

However, I’m having trouble getting the expected results. I always receive an "Invalid request" error, even though I’ve tried changing the requested fields and categories based on the documentation.

Could you please guide me on:

How to correctly request object detection and retrieve bounding box information (e.g., x, y, width, height) using Azure Content Understanding?

What fields and categories are required or supported to get object location data?

How can I troubleshoot or resolve the "Invalid request" error?

Any example requests, payload structures, or best practices would be greatly appreciated.

Thank you!