DermoSegDiff: A Boundary-aware Segmentation Diffusion Model for Skin Lesion Delineation

TL;DR

DermoSegDiff is a boundary-aware diffusion model that improves skin lesion segmentation by integrating boundary information and a specialized U-Net architecture, outperforming existing methods across multiple datasets.

Contribution

The paper introduces DermoSegDiff, a novel diffusion-based segmentation framework with a boundary-focused loss and a U-Net denoising network, advancing skin lesion delineation techniques.

Findings

Outperforms CNN, transformer, and diffusion models on skin segmentation datasets.

Effectively incorporates boundary information to enhance segmentation accuracy.

Demonstrates strong generalization across diverse skin lesion datasets.

Abstract

Skin lesion segmentation plays a critical role in the early detection and accurate diagnosis of dermatological conditions. Denoising Diffusion Probabilistic Models (DDPMs) have recently gained attention for their exceptional image-generation capabilities. Building on these advancements, we propose DermoSegDiff, a novel framework for skin lesion segmentation that incorporates boundary information during the learning process. Our approach introduces a novel loss function that prioritizes the boundaries during training, gradually reducing the significance of other regions. We also introduce a novel U-Net-based denoising network that proficiently integrates noise and semantic information inside the network. Experimental results on multiple skin segmentation datasets demonstrate the superiority of DermoSegDiff over existing CNN, transformer, and diffusion-based approaches, showcasing its effectiveness and generalization in various scenarios. The implementation is publicly accessible on \href{https://github.com/mindflow-institue/dermosegdiff}{GitHub}