Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

DFBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image Deepfake DetectionDFBench AI-Edited (test)
Object Enhance Accuracy57.32
9
Image Deepfake DetectionDFBench Overall (Full partition)
Accuracy54.44
9
Image Deepfake DetectionDFBench SD3.5 Large (AI-generated)
Accuracy36.41
8
Image Deepfake DetectionDFBench Playground AI-generated 1.0
Accuracy43.33
8
Showing 4 of 4 rows