Learning Cross-Image Object Semantic Relation in Transformer for Few-Shot Fine-Grained Image Classification
A Transformer-based double-helix model to achieve cross-image object semantic relation mining in a bidirectional and symmetrical manner.
bo-zhang