Image Generative Semantic Communication with Multi-Modal Similarity Estimation for Resource-Limited Networks.
Eri HosonumaTaku YamazakiTakumi MiyoshiAkihito TayaYuuki NishiyamaKaoru SezakiPublished in: CoRR (2024)
Keyphrases
- multi modal
- resource limited
- semantic concepts
- auto annotation
- input image
- uni modal
- image retrieval
- low level
- image features
- multi modality
- image content
- image representation
- image classification
- image analysis
- similarity estimation
- multiscale
- image segmentation
- single modality
- image annotation
- cross modal
- high dimensional
- segmentation algorithm
- low level features
- energy function
- high level