MAGIC: Multimodal relAtional Graph adversarIal inferenCe for Diverse and Unpaired Text-Based Image Captioning.
Wenqiao ZhangHaochen ShiJiannan GuoShengyu ZhangQingpeng CaiJuncheng LiSihui LuoYueting ZhuangPublished in: AAAI (2022)
Keyphrases
- input image
- relational graph
- image content
- multiscale
- image data
- image retrieval
- image features
- image representation
- image analysis
- high resolution
- image classification
- single image
- low level
- database
- segmentation method
- multimedia
- image segmentation
- feature points
- image collections
- random fields
- object recognition
- test images
- neural network