Improving Reinforcement Learning Based Image Captioning with Natural Language Prior.
Tszhang GuoShiyu ChangMo YuKun BaiPublished in: EMNLP (2018)
Keyphrases
- reinforcement learning
- natural language
- image data
- input image
- image representation
- template matching
- multiscale
- image analysis
- image segmentation
- image retrieval
- image regions
- image features
- hough transform
- low level
- prior information
- image classification
- image content
- vector field
- keypoints
- machine learning
- single image
- edge detection
- image collections
- similarity measure
- image set
- grey level
- prior model
- learning algorithm
- region of interest
- state space
- computer vision
- test images
- spatial information
- lighting conditions
- segmentation method
- bayesian framework
- gray level
- image structure
- image pixels
- high resolution
- image prior