Summary

In this chapter, we explored how image generation works. We also discussed the workings of multimodal models within Amazon Bedrock. We also covered several real-world applications and multimodal design patterns, including text-to-image, image search, image understanding, and image-to-image patterns such as inpainting and outpainting. We ended the chapter with a brief look at ethical considerations, as well as a look into the watermark detection capability within Amazon Bedrock. Throughout the chapter, we gained a deeper understanding of how we can leverage Amazon Bedrock’s multimodal models to build applications that can generate, understand, and manipulate images based on text and image prompts.

In the next chapter, we will explore the topic of building intelligent agents with Amazon Bedrock.