Unpaired Image Captioning by Image-level Weakly-Supervised Visual Concept Recognition.
Peipei ZhuXiao WangYong LuoZhenglong SunWei-Shi ZhengYaowei WangChangwen ChenPublished in: CoRR (2022)
Keyphrases
- image content
- image data
- image features
- input image
- image classification
- multiscale
- image collections
- segmentation method
- visual concepts
- weakly supervised
- image retrieval
- keypoints
- image representation
- segmentation algorithm
- spatial information
- partial occlusion
- feature extraction
- active learning
- pairwise
- superpixels
- keywords
- similarity measure