5 papers across 3 sessions
Self-supervised models learn robust representation of real images under cropping and resizing, which can be applied to detect AI-generated images without training.
A novel zero-shot generated image detection
This paper proposes a novel AIGI detection method based on Multi-granularity Local Entropy Patterns (MLEP), which captures scale- and location-invariant entropy features to improve accuracy and generalization across diverse generative models.