LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrieval.
Siqi SunYen-Chun ChenLinjie LiShuohang WangYuwei FangJingjing LiuPublished in: CoRR (2021)
Keyphrases
- text retrieval
- image retrieval
- object retrieval
- image content
- information retrieval
- low level
- image features
- text queries
- visual concepts
- visual features
- retrieval systems
- image classification
- image collections
- retrieval model
- bag of visual words
- latent semantic indexing
- web images
- text representation
- document collections
- image representation
- training set
- retrieval quality
- web search
- visually similar
- cross language
- document retrieval
- similarity measure
- medical image retrieval
- machine learning
- search engine
- inverted file
- visual content
- low level features
- vector space
- visual information
- image regions