Building Large Scale Image recommendation System using Vector Databases

AbstractWith the exponential growth of image data, efficiently retrieving similar images has become a critical challenge. This tutorial explores scalable image similarity search powered by vector databases and advanced deep learning models. Participants will gain insights into how Vision Foundation Models, such as Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), convert images into compact embeddings suitable for similarity comparison. We will delve into approximate nearest neighbor (ANN) search algorithms that enable fast and accurate image retrieval at scale. The tutorial further introduces modern vector database systems like Milvus and FAISS, which are designed to handle billion-scale image search tasks. Through practical demonstrations and architectural deep dives, attendees will learn how to build and deploy scalable, high-performance image similarity systems for real-world applications.

Organisers: Dr. Nasrullah (IBM), Dr. Sudipan Saha (IIT Delhi)