arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2604.00537cs.CVcs.AI
收藏

MATHENA:基于Mamba的牙科建筑层次估计器和整体评估网络

MATHENA: Mamba-based Architectural Tooth Hierarchical Estimator and Holistic Evaluation Network for Anatomy

OUTTA;Seoul National University(首尔国立大学);Sangmyung University(祥明大学);Gwangju Institute of Science and Technology(光州科学技术院);Samsung Medical Center, Sungkyunkwan University(三星医学中心,成均馆大学);Samsung Changwon Hospital, Sungkyunkwan University(三星昌原医院,成均馆大学);NVIDIA AI Technology Center(英伟达人工智能技术中心)
Kyeonghun Kim, Jaehyung Park, Youngung Han, Anna Jung, Seongbin Park, Sumin Lee, Jiwon Yang, Jiyoon Han, Subeen Lee, Junsu Lim, Hyunsu Go, Eunseob Choi, Hyeonse… 展开作者

Kyeonghun Kim, Jaehyung Park, Youngung Han, Anna Jung, Seongbin Park, Sumin Lee, Jiwon Yang, Jiyoon Han, Subeen Lee, Junsu Lim, Hyunsu Go, Eunseob Choi, Hyeonseok Jung, Soo Yong Kim, Woo Kyoung Jeong, Won Jae Lee, Pa Hong, Hyuk-Jae Lee, Ken Ying-Kai Liao, Nam-Joon Kim

AI总结:

MATHENA通过整合多分辨率SSM驱动检测器和轻量级Mamba-UNet,实现了牙科诊断的统一框架,包括龋齿分割、异常检测和牙科发育阶段分类,取得高精度结果。

摘要:

从正位全景片(OPGs)进行牙科诊断需要协调牙齿检测、龋齿分割(CarSeg)、异常检测(AD)和牙科发育阶段分类(DDS)。我们提出了MATHENA,一个利用Mamba线性复杂度状态空间模型(SSM)的统一框架,以解决所有四个任务。MATHENA整合MATHE,一个多分辨率SSM驱动检测器,包含四个方向的视觉状态空间(VSS)块,用于O(N)全局上下文建模,生成每颗牙齿的作物。这些作物由HENA处理,即轻量级Mamba-UNet,具有三重头架构和全局上下文状态标记(GCST)。在三重头架构中,CarSeg首先作为上游任务进行训练,以建立共享表示,然后冻结并重用于下游AD微调和DDS分类通过线性探测,实现稳定高效的学。我们还整理了PARTHENON基准,包含来自十个数据集的15,062个标注实例。MATHENA在牙齿检测中达到93.78%的mAP@50,在CarSeg中达到90.11%的Dice系数,在AD中达到88.35%,在DDS分类中达到72.40%的准确率。

英文摘要:

Dental diagnosis from Orthopantomograms (OPGs) requires coordination of tooth detection, caries segmentation (CarSeg), anomaly detection (AD), and dental developmental staging (DDS). We propose Mamba-based Architectural Tooth Hierarchical Estimator and Holistic Evaluation Network for Anatomy (MATHENA), a unified framework leveraging Mamba's linear-complexity State Space Models (SSM) to address all four tasks. MATHENA integrates MATHE, a multi-resolution SSM-driven detector with four-directional Vision State Space (VSS) blocks for O(N) global context modeling, generating per-tooth crops. These crops are processed by HENA, a lightweight Mamba-UNet with a triple-head architecture and Global Context State Token (GCST). In the triple-head architecture, CarSeg is first trained as an upstream task to establish shared representations, which are then frozen and reused for downstream AD fine-tuning and DDS classification via linear probing, enabling stable, efficient learning. We also curate PARTHENON, a benchmark comprising 15,062 annotated instances from ten datasets. MATHENA achieves 93.78% mAP@50 in tooth detection, 90.11% Dice for CarSeg, 88.35% for AD, and 72.40% ACC for DDS.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑