Built an efficient reverse image search application utilizing Pinecone's vector database and pretrained neural networks to generate vector embeddings.
- 10k+ images reverse image search in less than 100ms.
- Encoding images as vectors using a pretrained neural network (Imagenet)
- Conducting queries on Pinecone's Vector Database to identify images with the highest similarity scores to a given query image
Live Demo - https://visualsearch.streamlit.app/