Automated Image Object Detection and Indexing Workflow

This n8n workflow automates the process of analyzing an image to detect objects, crop those objects into individual images, and index the results for enhanced image search capabilities. Starting with a manual trigger, the workflow fetches a specified source image, uses Cloudflare’s AI service with the Detr-Resnet-50 model to identify objects within the image, filters results with high confidence scores (above 0.9), and crops out individual objects based on bounding box data. The cropped images are then uploaded to Cloudinary for storage, and relevant metadata along with image URLs are indexed in Elasticsearch. This setup allows for granular, object-based image search and retrieval, useful in scenarios such as e-commerce product catalogs, digital asset management, or content moderation systems.