VersaT2I: Improving Text-to-Image Models with Versatile Reward.
Jianshu GuoWenhao ChaiJie DengHsiang-Wei HuangTian YeYichen XuJiawei ZhangJenq-Neng HwangGaoang WangPublished in: CoRR (2024)
Keyphrases
- image data
- image features
- input image
- random fields
- image retrieval
- image content
- parametric models
- image representation
- image classification
- image analysis
- image statistics
- feature points
- image collections
- template matching
- single image
- edge detection
- probabilistic model
- information retrieval
- statistical model
- image pixels
- multiscale
- similarity measure
- bayesian framework
- object models
- image matching
- high resolution
- pixel values
- image regions
- text information
- text graphics
- region of interest
- test images
- spatial information
- hough transform
- segmentation algorithm
- text mining
- denoising
- image segmentation