Topics

FMCerNet analyzes larger image regions and overlapping abnormalities, improving cervical cell screening and subtype classification in tests on public and private datasets.

Complementary detectors improve results, but deployment conditions still matter.

Deepfake detection turns to light: a hybrid optical system screened 15 videos simultaneously with 97.79% accuracy in laboratory tests, raising prospects for less costly synthetic media screening.