Multi-defect microscopy image restoration under limited data conditions
This addresses the challenge of limited data for researchers using deep learning in microscopy image restoration, though it is incremental as it builds on existing GAN techniques.
The authors tackled the problem of restoring multi-defect fluorescence microscopy images under limited training data by proposing a two-stage method using GANs, achieving comparable or better results than existing methods like CARE, deblurGAN, and CycleGAN in image quality.
Deep learning methods are becoming widely used for restoration of defects associated with fluorescence microscopy imaging. One of the major challenges in application of such methods is the availability of training data. In this work, we propose a unified method for reconstruction of multi-defect fluorescence microscopy images when training data is limited. Our approach consists of two stages: first, we perform data augmentation using Generative Adversarial Network (GAN) with conditional instance normalization (CIN); second, we train a conditional GAN (cGAN) on paired ground-truth and defected images to perform restoration. The experiments on three common types of imaging defects with different amounts of training data show that the proposed method gives comparable results or outperforms CARE, deblurGAN and CycleGAN in restored image quality when available data is limited.