CgT-GAN: CLIP-guided Text GAN for Image Captioning.
Jiarui YuHaoran LiYanbin HaoBin ZhuTong XuXiangnan HePublished in: CoRR (2023)
Keyphrases
- image classification
- image content
- multiscale
- image analysis
- template matching
- single image
- image data
- input image
- image retrieval
- image features
- segmentation method
- image pixels
- image segmentation
- test images
- image collections
- low level
- web images
- text mining
- image structure
- high resolution
- image regions
- information retrieval
- video clips
- complex background
- image set
- textual descriptions
- text information
- region of interest
- image matching
- keypoints
- segmentation algorithm
- image representation
- visual features
- edge detection
- color images
- similarity measure
- high level