IEEE Trans Med Imaging. 2021 Aug;40(8):2042-2052. doi: 10.1109/TMI.2021.3070847. Epub 2021 Jul 30.
Instance level detection and segmentation of thoracic diseases or abnormalities are crucial for automatic diagnosis in chest X-ray images. Leveraging on constant structure and disease relations extracted from domain knowledge, we propose a structure-aware relation network (SAR-Net) extending Mask R-CNN. The SAR-Net consists of three relation modules: 1. the anatomical structure relation module encoding spatial relations between diseases and anatomical parts. 2. the contextual relation module aggregating clues based on query-key pair of disease RoI and lung fields. 3. the disease relation module propagating co-occurrence and causal relations into disease proposals. Towards making a practical system, we also provide ChestX-Det, a chest X-Ray dataset with instance-level annotations (boxes and masks). ChestX-Det is a subset of the public dataset NIH ChestX-ray14. It contains ~3500 images of 13 common disease categories labeled by three board-certified radiologists. We evaluate our SAR-Net on it and another dataset DR-Private. Experimental results show that it can enhance the strong baseline of Mask R-CNN with significant improvements. The ChestX-Det is released at https://github.com/Deepwise-AILab/ChestX-Det-Dataset.
胸部疾病或异常的实例级检测和分割对于胸部 X 光图像的自动诊断至关重要。利用从领域知识中提取的恒定结构和疾病关系，我们提出了一种基于结构感知关系网络（SAR-Net）扩展 Mask R-CNN。SAR-Net 由三个关系模块组成：1. 解剖结构关系模块，用于编码疾病和解剖部位之间的空间关系。2. 上下文关系模块，基于疾病 RoI 和肺区的查询-键对聚合线索。3. 疾病关系模块，将共同出现和因果关系传播到疾病提案中。为了构建一个实用的系统，我们还提供了 ChestX-Det，这是一个具有实例级注释（框和掩码）的胸部 X 射线数据集。Chestx-Det 是公共数据集 NIH ChestX-ray14 的一个子集。它包含由三位认证放射科医生标记的 13 种常见疾病类别的约 3500 张图像。我们在它和另一个数据集 DR-Private 上评估了我们的 SAR-Net。实验结果表明，它可以通过显著的改进来增强 Mask R-CNN 的强大基线。Chestx-Det 可在 https://github.com/Deepwise-AILab/ChestX-Det-Dataset 上获得。
