Downstream Task-agnostic Transferable Attacks on Language-Image Pre-training Models.
Yiqiang LvJingjing ChenZhipeng WeiKai ChenZuxuan WuYu-Gang JiangPublished in: ICME (2023)
Keyphrases
- multiscale
- image data
- image content
- random fields
- input image
- image segmentation
- image features
- image retrieval
- image statistics
- image analysis
- structured prediction
- image pixels
- single image
- image collections
- high resolution
- object models
- parametric models
- feature points
- natural language
- bayesian framework
- watermarking technique
- similarity measure
- test images
- statistical model
- image regions
- image representation
- image classification
- edge detection
- programming language
- low level
- countermeasures
- supply chain
- probabilistic model
- visual effects