Computer Vision on AWS

12

Best Practices for Designing an End-to-End CV Pipeline

In the previous chapter, we introduced Amazon Augmented AI (A2I). We discussed the importance of a human-in-the-loop workflow and walked through a code example while using Amazon Rekognition to analyze unsafe images.

Throughout this book, we’ve covered several real-world CV use cases. We’ve also discussed AWS AI/ML services in detail, including Rekognition, Lookout for Vision, and SageMaker. To recap, AWS AI/ML services are managed services, which means that the undifferentiated heavy lifting of patching, upgrading, and maintaining servers and hardware is removed. AWS AI services are composed of serverless architecture that scales automatically. Rekognition includes pre-trained capabilities, ...