no code implementations • 2 Apr 2018 • Dongwook Lee, Jaejun Yoo, Sungho Tak, Jong Chul Ye
The proposed deep residual learning networks are composed of magnitude and phase networks that are separately trained.