用对抗生成与条件随机场提升脑肿瘤三维分割精度
Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field
- 结合V-net与条件随机场，引入伪3D结构增强细节捕捉
- 在BraTS-2018数据集上特异性超99.8%，优于U-net等经典模型
- 适合需要高精度医学图像分割的研究者或临床辅助场景
由于胶质瘤结构复杂且个体差异大，精准的脑肿瘤分割仍具挑战。本文提出一种多模态3D体积生成对抗网络（3D-vGAN），融合条件随机场（CRF）的强细节鲁棒性与V-net的空间特征提取能力。通过伪3D结构改进V-net，将条件随机场置于生成器后，并以原始图像作为补充引导。在BraTS-2018数据集上的实验表明，3D-vGAN显著优于U-net、GAN、FCN和3D V-net等经典模型，特异性超过99.8%。
原文摘要 · Abstract (English)
Accurate brain tumor segmentation remains a challenging task due to structural complexity and great individual differences of gliomas. Leveraging the pre-eminent detail resilience of CRF and spatial feature extraction capacity of V-net, we propose a multimodal 3D Volume Generative Adversarial Network (3D-vGAN) for precise segmentation. The model utilizes Pseudo-3D for V-net improvement, adds conditional random field after generator and use original image as supplemental guidance. Results, using the BraTS-2018 dataset, show that 3D-vGAN outperforms classical segmentation models, including U-net, Gan, FCN and 3D V-net, reaching specificity over 99.8%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。