Stain Style Transfer Using Transitive Adversarial Networks

Abstract

Digitized pathological diagnosis has been in increasing demand recently. It is well known that color information is critical to the automatic and visual analysis of pathological slides. However, the color variations due to various factors not only have negative impact on pathologist's diagnosis, but also will reduce the robustness of the algorithms. The factors that cause the color differences are not only in the process of making the slices, but also in the process of digitization. Different strategies have been proposed to alleviate the color variations. Most of such techniques rely on collecting color statistics to perform color matching across images and highly dependent on a reference template slide. Since the pathological slides between hospitals are usually unpaired, these methods do not yield good matching results. In this work, we propose a novel network that we refer to as Transitive Adversarial Networks (TAN) to transfer the color information among slides from different hospitals or centers. It is not necessary for an expert to pick a representative reference slide in the proposed TAN method. We compare the proposed method with the state-of-the-art methods quantitatively and qualitatively. Compared with the state-of-the-art methods, our method yields an improvement of 0.87dB in terms of PSNR, demonstrating the effectiveness of the proposed TAN method in stain style transfer.

Generative Adversarial Networks in Digital Pathology and Histopathological Image Processing: A Review

“…The listed color normalization approaches are based on a style transfer method in which the style of the input image is modified based on the style image, when preserving the content of the input image. [ 37 39 50 51 53 54 55 ] The methods based on cycleGAN explore the capability of unpaired image-to-image translation which makes it a flexible architecture for stain normalization. Other approaches discussed here use alternative formulations such as self-attention models,[ 56 ] cGAN,[ 31 ] and encoder–decoder architecture.…”

StainNet: A Fast and Robust Stain Normalization Network

“…However, the potential factor in the staining reagent, staining process, and slide scanner specifications often result in inconsistency of pathological images (1). These variations not only affect the judgment of pathologists but also weaken the performance of CAD systems and hamper their applications in pathology (2)(3)(4). So, stain normalization is a routine pre-processing operation for pathological images, especially for CAD systems, and it is reported to help increase the prediction accuracy, such as tumor classification (5).…”

“…Deep learning-based methods mostly apply generative adversarial networks (GANs) to achieve stain normalization (3,7,8,(16)(17)(18). Shaban et al (8) proposed an unsupervised stain normalization method named StainGAN based on CycleGAN (16) to transfer the stain style.…”

“…Shaban et al (8) proposed an unsupervised stain normalization method named StainGAN based on CycleGAN (16) to transfer the stain style. Cai et al (3) proposed a new generator to improve the image quality and accelerate the networks. On the other hand, Cho et al (18), Salehi et al (7), and Tellez et al (17) reconstructed original images from the images with color augmentations, e.g., grayscale and Hue-Saturation-Value (HSV) transformation, and tried to normalize other color styles to the original.…”

“…However, inconsistencies in staining process, operator ability, and scanner specifications often result in different appearances of pathological images. These variations would weaken the performance of CAD systems and hamper their applications in pathology [1][2][3]. Currently, stain normalization algorithms have been proposed to reduce color variation in stained images.…”