The correct answer is Detect objects and scenes in images. Amazon Rekognition provides built-in label detection that can identify objects scenes and activities in images immediately without requiring you to build custom models.

Rekognition’s label detection returns categories and confidence scores for items it recognizes so you can quickly catalog or filter visual content. The service also includes ready-made features for face detection face analysis text-in-image moderation and PPE detection which makes it suitable for common vision tasks out of the box.

Convert speech in videos to text is incorrect because that capability belongs to Amazon Transcribe rather than Rekognition.

Automatically resize or compress images is incorrect since Rekognition performs visual analysis and not image editing or transformation.

Estimate full-body human pose is incorrect because Rekognition does not offer a general full-body pose estimation feature and pose analysis is typically handled with specialized models or other tools.

Full AWS Practitioner Certification Question