[P-IPDP] ISOM'22 Poster Session (Post Deadline)

[P-IPDP-01] A Deep Neural Network for Time-Fluctuation Spatial Noise Suppression in Optical Correlation Imaging based on a Progressive Growing Adversarial Network

*Hiroki Takahara¹, Yuta Wada¹, Kaito Nakao¹, Taku Hosizawa¹, Eriko Watanabe¹ (1. The University of Electro-Communications (Japan))

[Presentation Style] Onsite

We propose a deep neural network for time-fluctuation noise suppression (TFNS) by applying a progressive growing GAN (PG-GAN), called a PG-TFNS network. The accuracy of the proposed approach was improved by using an efficient sub-pixel convolutional neural network (ESPCN) for its super-resolution part.