BDNet: A BERT-based dual-path network for text-to-image cross-modal person re-identification.
Qiang LiuXiaohai HeQizhi TengLinbo QingHonggang ChenPublished in: Pattern Recognit. (2023)
Keyphrases
- cross modal
- person re identification
- input image
- image data
- image retrieval
- image classification
- image features
- image representation
- image content
- image segmentation
- visual similarity
- multi modal
- multiscale
- low level
- image collections
- feature representation
- similarity measure
- web images
- spatial information
- object recognition
- visual recognition
- camera network