ELARKHNU browse.title.page

Зараз показуємо 1 - 1 з 1

Semantic search of relevant images using vector databases
(2025) Ostapchenko, N.; Tyschenko, O.; Denysenko, B.; Mazurets, O.; Мазурець, Олександр Вікторович
This paper presents a method for semantic image search using vector databases, enabling efficient retrieval of relevant visual content based on textual or visual queries. The proposed approach leverages state-of-the-art embedding models, such as OpenCLIP, to convert both images and queries into high-dimensional vector representations. These vectors are stored and compared in a vector database to determine semantic similarity. The system supports both text-to-image and image-to-image search, significantly enhancing the precision of results by focusing on content-level meaning rather than keywords. The solution is applicable in various domains requiring fast and accurate access to large-scale image repositories.