
Amazon Kendra combined with generative AI automatically generates image captions and metadata to enable enterprise-scale searchable image repositories without manual tagging effort.


<div><p>This article demonstrates how to automate image captioning and enable searchable image repositories using generative AI and Amazon Kendra.</p><ul><li>Generative AI models automatically create textual captions for images, eliminating manual metadata tagging</li><li>Amazon Kendra's Custom Document Enrichment invokes AI models during image ingestion to generate searchable metadata</li><li>Solution combines vision and language models (ViT-GPT2) deployed on Amazon SageMaker with Amazon Textract</li><li>Applicable across ecommerce, healthcare, manufacturing, marketing, and accessibility use cases</li><li>CloudFormation template provided for easy deployment with estimated monthly cost of $129.20</li><li>Users can search images using natural language queries like "dog under umbrella"</li></ul><p>This solution automates enterprise-scale image search by leveraging generative AI to create rich, searchable metadata without manual effort.</p></div>


Related articles