Adversarial Stain Transfer to Study the Effect of Color Variation on Cell Instance Segmentation
This work addresses the challenge of color variation for pathologists and segmentation algorithms in immunohistochemical staining images, but it is incremental as it quantifies an existing issue rather than introducing a new method.
The study tackled the problem of stain color variation in histological images affecting cell segmentation by producing NeuN staining images with different colors and using adversarial stain transfer to quantify the impact. The results showed that color normalization is necessary for accurate segmentation, with performance drops of up to 15% in F1-score without it.
Stain color variation in histological images, caused by a variety of factors, is a challenge not only for the visual diagnosis of pathologists but also for cell segmentation algorithms. To eliminate the color variation, many stain normalization approaches have been proposed. However, most were designed for hematoxylin and eosin staining images and performed poorly on immunohistochemical staining images. Current cell segmentation methods systematically apply stain normalization as a preprocessing step, but the impact brought by color variation has not been quantitatively investigated yet. In this paper, we produced five groups of NeuN staining images with different colors. We applied a deep learning image-recoloring method to perform color transfer between histological image groups. Finally, we altered the color of a segmentation set and quantified the impact of color variation on cell segmentation. The results demonstrated the necessity of color normalization prior to subsequent analysis.