Visual Spatial Description: Controlled Spatial-Oriented Image-to-Text Generation

31472

Knowledge Triples

Source Entity	Relationship	Target Entity
Joint End to End Architecture Method	improves	Visual Spatial Description Task
Joint End to End Architecture Method	improves	Visual Spatial Relationship Classification Task
Visual Spatial Description Task	integrates	Visual Spatial Relationship Classification Task
Visual Spatial Description Task	uses	Benchmark Dataset Material