NUS-WIDE

The NUS-WIDE dataset is a large-scale benchmark that contains images with multiple labels, used to evaluate multi-label image recognition tasks, particularly in scenarios involving missing labels.

Papers using NUS-WIDE (3)

PVLR: Prompt-driven Visual-Linguistic Representation Learning for Multi-Label Image Recognition2024 · 1 citesRethinking Sparse Lexical Representations for Image Retrieval in the Age of Rising Multi-Modal Large Language Models2024 · 1 citesText-Region Matching for Multi-Label Image Recognition with Missing Labels2024