logo-similarity
rghafadaryan/logo-similarity
Detector of Brand Similarity for Armenian Language
computer-visionimage-similaritylogo-detectionvector-searchfaissembeddingsperceptual-hashingmultimodal-search
Summary
An experimental toolkit for logo and image similarity detection that combines multiple computer vision techniques: AI embeddings (DINO, CLIP), perceptual hashing (pHash), and Fourier shape descriptors. It builds a local searchable database using FAISS for vector similarity search and SQLite for metadata, with command-line scripts for adding images and querying matches.