MedSegDiffNCA: Diffusion Models With Neural Cellular Automata for Skin Lesion Segmentation

2025-01-05

AI Summary

To address the high computational cost and poor deployability of diffusion models—typically reliant on large U-Net backbones—in resource-constrained medical settings for skin lesion segmentation, this work proposes the first lightweight diffusion-based segmentation framework built upon Neural Cellular Automata (NCA). Methodologically, we introduce a multi-level NCA noise refinement mechanism, an attention-enhanced NCA architecture incorporating CBAM, and an RGB-channel semantic-guided loss, enabling the first synergistic optimization of NCA with Denoising Diffusion Probabilistic Models (DDPM) for medical image segmentation. Experiments demonstrate that our approach achieves a Dice score of 87.84% on skin lesion segmentation—on par with U-Net-based baselines—while reducing model parameters by 60–110×. This substantial parameter reduction significantly enhances feasibility for deployment in low-resource clinical environments.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Learning to Search

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Denoising Diffusion Models (DDMs) are widely used for high-quality image generation and medical image segmentation but often rely on Unet-based architectures, leading to high computational overhead, especially with high-resolution images. This work proposes three NCA-based improvements for diffusion-based medical image segmentation. First, Multi-MedSegDiffNCA uses a multilevel NCA framework to refine rough noise estimates generated by lower level NCA models. Second, CBAM-MedSegDiffNCA incorporates channel and spatial attention for improved segmentation. Third, MultiCBAM-MedSegDiffNCA combines these methods with a new RGB channel loss for semantic guidance. Evaluations on Lesion segmentation show that MultiCBAM-MedSegDiffNCA matches Unet-based model performance with dice score of 87.84% while using 60-110 times fewer parameters, offering a more efficient solution for low resource medical settings.

Problem

• Medical Image Segmentation
• Denoising Diffusion Models
• Efficient Unet Architecture

Innovation

• MultiCBAM-MedSegDiffNCA
• Efficient Medical Image Segmentation
• Noisy Diffusion Model