Azure AI Vision Image Analysis: captions, dense captions, tags, objects, people and smart crops

Azure AI Vision's Image Analysis API describes what is in an image using prebuilt models trained by Microsoft, with no training on your part. You send an image (as a URL or as binary data) and a list of visual features you want, and the service returns JSON with results and confidence scores. The current version, Image Analysis 4.0, is called at the imageanalysis:analyze path with a features query parameter, or through the ImageAnalysisClient in the Python and C# SDKs.