Focus on Content not Noise: Improving Image Generation for Nuclei Segmentation by Suppressing Steganography in CycleGAN
This work addresses the laborious and variable task of annotating nuclei in microscopy images for training neural networks, though it is incremental as it builds on existing CycleGAN methods.
The paper tackled the problem of content inconsistencies in synthetic microscopy images generated by CycleGAN for nuclei segmentation by suppressing steganography using low-pass filtering based on DCT, resulting in a 5.4 percentage point improvement in F1-score compared to vanilla CycleGAN.
Annotating nuclei in microscopy images for the training of neural networks is a laborious task that requires expert knowledge and suffers from inter- and intra-rater variability, especially in fluorescence microscopy. Generative networks such as CycleGAN can inverse the process and generate synthetic microscopy images for a given mask, thereby building a synthetic dataset. However, past works report content inconsistencies between the mask and generated image, partially due to CycleGAN minimizing its loss by hiding shortcut information for the image reconstruction in high frequencies rather than encoding the desired image content and learning the target task. In this work, we propose to remove the hidden shortcut information, called steganography, from generated images by employing a low pass filtering based on the DCT. We show that this increases coherence between generated images and cycled masks and evaluate synthetic datasets on a downstream nuclei segmentation task. Here we achieve an improvement of 5.4 percentage points in the F1-score compared to a vanilla CycleGAN. Integrating advanced regularization techniques into the CycleGAN architecture may help mitigate steganography-related issues and produce more accurate synthetic datasets for nuclei segmentation.