Great video! Embarking on an incredible journey of vector similarity search with k-nearest neighbor (k-NN) in Postgres queries is truly fascinating. Generating vector embeddings of facial images using the deepface package for Python and conducting lightning-fast k-NN searches opens up a world of possibilities in image analysis, reverse image search, recommendation engines, and NLP applications like ChatGPT. Count me in to revolutionize my search capabilities with k-NN and Postgres! Thank you for sharing this amazing content!
Great video, if I may ask, what does pgverctor library for python adds on to your method? I'm looking for lowest latency for similarity search, and I think your approach is faster and straight to the point for using postgress to store embeddings and compare similarities, or Im missing something? Thank you
First I wanna say thanks a lot for your research mate, I got a question is it possible for me to create an enrollment interface that I can use later so I assign privileges and permissions depending on the real time face of a person?
hi sir sefik, in terms of speed and precision, what is the most compliant database with deep face for vector comparison, do you think faiss data base is the best ?
@@sefiks Wouldn't it be better to store your embeddings in to a vector database instead since we're working with vectors? I've recently heard about them so I don't know much, but it seems reasonable you'd benefit from storing your vectors in a vector database, no?