Duplicate and Near-Duplicate Image Detection for Digital Asset Libraries

Entreprise partenaire — révélée une fois votre stage confirmé

Compétences recherchées

Sujet du stage

Find duplicate and near-duplicate images across a growing library — a project combining embeddings, similarity search, and real engineering trade-offs at scale. What you'll work on : • Build an image-hashing or embedding-based pipeline to detect duplicate/near-duplicate images • Benchmark detection accuracy on a test set of known duplicates • Store and query embeddings efficiently for a growing image library What we're looking for : • Academic or hands-on experience with React/Next.js is a plus. • Familiarity with Image Hashing/CLIP Embeddings is a plus. • Initiative and a sharp eye for detail. What you'll get : • A dedicated computer-vision technical mentor, with follow-up based on your progress report • A concrete product-facing use case • Completion certificate based on how far you get • Most engaged candidates may be recommended for pre-hire or job opportunities