ABOUT

PEOPLE

ADMISSION

EDUCATION

ARCHIVE

BOARD

  • ARCHIVE

MV CIMF Festival

MoCHA-former: Moiré-Conditioned Hybrid Adaptive Transformer for Video Demoiréing

Jeahun Sung, Changhyun Roh, Chanho Eom, Jihyong Oh Creative Vision and Multimedia Lab

Recent advances in portable imaging have made screen capture common, but frequency aliasing between the camera CFA and display sub-pixels causes severe moiré artifacts. We propose MoCHA-former, a video demoiréing framework with two key modules: DMAD, which separates moiré from content to produce moiré-adaptive features, and STAD, which captures large-scale structures, models channel dependence, and ensures temporal consistency without explicit alignment. Evaluations on RAW and sRGB video datasets show that MoCHA-former achieves superior PSNR, SSIM, and LPIPS compared to prior methods.