Multi-style conversion for semantic segmentation of lesions in fundus images by adversarial attacks

TL;DR

This paper presents an adversarial style conversion method that enables a single segmentation model to adapt to different annotation styles across multiple fundus image datasets, improving generalization and interpretability in diabetic retinopathy diagnosis.

Contribution

The novel adversarial style conversion technique allows a unified model to handle diverse annotation styles without needing separate models or extensive dataset standardization.

Findings

Model successfully converts segmentation styles across datasets.

Enhanced generalization and uncertainty estimation demonstrated.

Facilitates continuous interpolation between annotation styles.

Abstract

The diagnosis of diabetic retinopathy, which relies on fundus images, faces challenges in achieving transparency and interpretability when using a global classification approach. However, segmentation-based databases are significantly more expensive to acquire and combining them is often problematic. This paper introduces a novel method, termed adversarial style conversion, to address the lack of standardization in annotation styles across diverse databases. By training a single architecture on combined databases, the model spontaneously modifies its segmentation style depending on the input, demonstrating the ability to convert among different labeling styles. The proposed methodology adds a linear probe to detect dataset origin based on encoder features and employs adversarial attacks to condition the model's segmentation style. Results indicate significant qualitative and quantitative through dataset combination, offering avenues for improved model generalization, uncertainty estimation and continuous interpolation between annotation styles. Our approach enables training a segmentation model with diverse databases while controlling and leveraging annotation styles for improved retinopathy diagnosis.