Posted On: Dec 20, 2022
Amazon Rekognition Labels is a machine learning-based image and video analysis service that can detect objects, people, text, scenes, and activities. On 11/23/2022, we launched an update to the machine learning model of Amazon Rekognition Labels Image to add 600 labels and improve the accuracy of over 2,000 existing labels. We also added an ability to organize label results by “aliases” and “categories” and support filtering of results.
Starting today, these improvements are available for stored video analysis. In addition, we now optionally return timestamps and duration of the video segments for when a label is detected across multiple frames in a video. These updates are now available in all AWS regions supported for Amazon Rekognition Labels Video APIs. To learn more, read our blog and the Amazon Rekognition Labels documentation.