[论文解读] A Mutual Bootstrapping Model for Automated Skin Lesion Segmentation and Classification

Yutong Xie, Jianpeng Zhang+2 more|arXiv (Cornell University)|Mar 8, 2019

Cutaneous Melanoma Detection and Management参考文献 59被引用 14

一句话总结

该论文提出了一种互惠自举深度学习模型（MB-DCNN），通过实现双向知识迁移，联合执行皮肤病变分割与分类：粗分割掩码指导分类中的病变定位，而分类的定位图则提升分割精度。该方法实现了最先进性能，在ISIC-2017数据集上的Jaccard指数为80.4%，在PH2数据集上为89.4%，分类任务的平均AUC分别为93.8%和97.7%。

ABSTRACT

Automated skin lesion segmentation and classification are two most essential and related tasks in the computer-aided diagnosis of skin cancer. Despite their prevalence, deep learning models are usually designed for only one task, ignoring the potential benefits in jointly performing both tasks. In this paper, we propose the mutual bootstrapping deep convolutional neural networks (MB-DCNN) model for simultaneous skin lesion segmentation and classification. This model consists of a coarse segmentation network (coarse-SN), a mask-guided classification network (mask-CN), and an enhanced segmentation network (enhanced-SN). On one hand, the coarse-SN generates coarse lesion masks that provide a prior bootstrapping for mask-CN to help it locate and classify skin lesions accurately. On the other hand, the lesion localization maps produced by mask-CN are then fed into enhanced-SN, aiming to transfer the localization information learned by mask-CN to enhanced-SN for accurate lesion segmentation. In this way, both segmentation and classification networks mutually transfer knowledge between each other and facilitate each other in a bootstrapping way. Meanwhile, we also design a novel rank loss and jointly use it with the Dice loss in segmentation networks to address the issues caused by class imbalance and hard-easy pixel imbalance. We evaluate the proposed MB-DCNN model on the ISIC-2017 and PH2 datasets, and achieve a Jaccard index of 80.4% and 89.4% in skin lesion segmentation and an average AUC of 93.8% and 97.7% in skin lesion classification, which are superior to the performance of representative state-of-the-art skin lesion segmentation and classification methods. Our results suggest that it is possible to boost the performance of skin lesion segmentation and classification simultaneously via training a unified model to perform both tasks in a mutual bootstrapping way.

研究动机与目标

• 解决现有深度学习模型将皮肤病变分割与分类视为孤立任务的局限性。
• 利用分割与分类任务之间的互补性，同时提升两个任务的性能。
• 通过弱监督知识迁移利用图像级标签，减少对昂贵像素级标注的依赖。
• 提出一种新型训练策略，实现在分割与分类网络之间的双向知识迁移。
• 通过混合损失函数提升对类别不平衡及病变分割中难易像素不平衡的鲁棒性。

提出的方法

• 提出三阶段网络架构：粗粒度-SN用于初始病变掩码预测，mask-CN用于注意力引导的分类，enhanced-SN用于精细化分割。
• 利用粗粒度-SN的预测结果作为先验知识，指导mask-CN学习更准确的病变定位与分类。
• 将mask-CN的定位图反向传输至enhanced-SN以优化分割，实现互惠自举。
• 引入一种新型排序损失，结合Dice损失，以缓解分割任务中的类别不平衡与难易像素不平衡问题。
• 使用ISIC-2017和PH2数据集中的图像级（分类）与像素级（分割）标注，端到端训练模型。
• 通过混合损失函数优化联合训练过程，平衡分割精度与泛化能力。

实验结果

研究问题

• RQ1 与单任务模型相比，联合训练分割与分类是否能提升两个任务的性能？
• RQ2 分割与分类网络之间的双向知识迁移在提升模型精度方面有多有效？
• RQ3 在弱监督学习框架下，图像级标注在多大程度上能提升分割性能？
• RQ4 所提出的混合损失函数（Dice损失 + 排名损失）是否能有效缓解病变分割中的类别不平衡与难易像素不平衡问题？
• RQ5 在标注数据有限的情况下，互惠自举机制的表现如何？

主要发现

• MB-DCNN模型在ISIC-2017数据集上的皮肤病变分割Jaccard指数达到80.4%，在PH2数据集上达到89.4%，优于现有最先进方法。
• 该模型在ISIC-2017数据集上的分类平均AUC为93.8%，在PH2数据集上为97.7%，超过现有最先进模型的性能。
• 互惠自举机制显著提升了性能，尤其在训练数据有限时表现更优。
• 将排名损失与Dice损失结合，有效减轻了类别不平衡及难易像素不平衡对分割的影响。
• 模型展现出强大的泛化能力，在包含更新、更多样化的皮肤镜图像的ISIC数据集归档中仍保持高性能。
• 在线推理效率高，单张图像在GPU服务器上处理时间不足1秒，表明具备临床部署可行性。