SIAN: Style-Guided Instance-Adaptive Normalization for Multi-Organ Histopathology Image Synthesis

TL;DR

This paper introduces SIAN, a novel normalization method that synthesizes realistic multi-organ histopathology images with accurate nuclei boundaries, improving image realism and segmentation performance.

Contribution

SIAN is a new style-guided normalization approach that enhances multi-organ histopathology image synthesis and boundary accuracy, surpassing existing methods.

Findings

Outperforms four state-of-the-art methods in realism across five organs.

Enables a segmentation network to achieve state-of-the-art results using synthetic images.

Generates more accurate nuclei boundaries and diverse styles.

Abstract

Existing deep neural networks for histopathology image synthesis cannot generate image styles that align with different organs, and cannot produce accurate boundaries of clustered nuclei. To address these issues, we propose a style-guided instance-adaptive normalization (SIAN) approach to synthesize realistic color distributions and textures for histopathology images from different organs. SIAN contains four phases, semantization, stylization, instantiation, and modulation. The first two phases synthesize image semantics and styles by using semantic maps and learned image style vectors. The instantiation module integrates geometrical and topological information and generates accurate nuclei boundaries. We validate the proposed approach on a multiple-organ dataset, Extensive experimental results demonstrate that the proposed method generates more realistic histopathology images than four state-of-the-art approaches for five organs. By incorporating synthetic images from the proposed approach to model training, an instance segmentation network can achieve state-of-the-art performance.