Incorporating Copying Mechanism in Image Captioning for Learning Novel Objects.
Ting YaoYingwei PanYehao LiTao MeiPublished in: CVPR (2017)
Keyphrases
- input image
- image pixels
- reinforcement learning
- image features
- image data
- single image
- learning process
- image representation
- image analysis
- image content
- image segmentation
- spatial information
- image regions
- complex scenes
- relative position
- segmentation method
- unlabeled images
- image classification
- multiscale
- low level
- image retrieval
- spatial relations
- image processing
- learning mechanism
- similarity measure
- scene understanding
- bounding box
- moving objects
- spatial relationships
- e learning
- high resolution
- keypoints
- d objects
- image restoration
- supervised learning
- edge detection