Quality Assessment of Synthetic Fluorescence Microscopy Images for Image Segmentation

Synthetic images are widely used in image segmentation for algorithm training and performance assessment. Recently, advances in image synthesis techniques, especially generative adversarial networks (GANs), have made it possible to generate fluorescence microscopy images with remarkably realistic appearance. However, intuitive and specific metrics to assess the quality of these images remain lacking. Here, we propose three quality metrics that quantify the fidelity of the foreground signal, the background noise, and blurring, respectively, of synthesized fluorescence microscopy images. Using these metrics, we examine images of mitochondria synthesized by two representative GANs: pix2pix, which requires paired training data, and CycleGAN, which does not require paired training data. We find that both networks generate realistic images and achieve similar fidelity in reproducing background noise and blurring of real images. However, CycleGAN achieves significantly higher fidelity than pix2pix in reproducing intensity patterns of real mitochondria. When used to train the U-Net for segmentation, images synthesized by both networks achieve performance on par with real images. Overall, we have developed a method to assess quality of synthetic fluorescence microscopy images and to evaluate their training performance in image segmentation. The quality metrics proposed are general and can be used to assess fluorescence microscopy images synthesized by different methods.

1. INTRODUCTION

Synthetic images are used extensively in computational analysis of biological images in applications such as image segmentation [1] and feature tracking [2]. In image segmentation, synthetic images are often used for training and evaluating algorithms because their ground truth is known and, therefore, requires no manual labelling. Another advantage of synthetic images is that their conditions such as signal-to-noise ratios (SNRs) often can be controlled [3], [4]. Despite these advantages, performance of synthetic images is fundamentally defined by the level of fidelity they achieve in mimicking real images.