Presentation

<
73. Radial Residual Frequency: A Semantically Aligned Benchmark and Spectral Detector for AI-Generated Images
SessionPosters: Images, Video & Computer Vision
DescriptionWe release a semantically aligned real/synthetic image benchmark and Radial Residual Frequency (rrF), a lightweight spectral metric that captures directional asymmetry in the FFT log-magnitude. A ResNet-50+rrF detector reaches 99.58% accuracy on 4,000 paired images across SDXL, Stable Diffusion 3.5, and GPT-Image 1, substantially outperforming existing detectors.
Contributors
Guang Yang
University of California, Berkeley
Qian Zhang
College of Charleston
Fengchen Liu
University of California, Berkeley
Event Type
Poster
TimeSunday, 19 July 20269:00am - 5:30pm PDT
LocationWest Hall Lobby
Session TimeSunday, 19 July 20269:00am - 5:30pm PDTMonday, 20 July 20269:00am - 5:30pm PDTTuesday, 21 July 20269:00am - 5:30pm PDTWednesday, 22 July 20269:00am - 5:30pm PDTThursday, 23 July 20269:00am - 5:30pm PDT
LocationWest Hall Lobby
Registration Categories
Full Conference Supporter
Full Conference
Experience
Next PresentationNext Presentation74. ResonantLoom: Neural Acoustic Rendering for Editable Multi-Entity Video-to-Audio Synthesis