Scholarlink

<p>Particle identification (PID) is essential for future particle physics experiments such as the Circular Electron-Positron Collider (CEPC) and the Future Circular Collider. A high-granularity Time Projection Chamber (TPC) not only provides precise tracking but also enables d<i>N</i>/d<i>x</i> measurements for PID. The d<i>N</i>/d<i>x</i> method estimates the number of primary ionization electrons, offering significant improvements in PID performance. However, accurate reconstruction remains a major challenge for this approach. In this paper, we introduce a deep learning model, the Graph Point Transformer (GraphPT), for d<i>N</i>/d<i>x</i> reconstruction. In our approach, TPC data are represented as point clouds. The network backbone adopts a U-Net architecture built upon graph neural networks, incorporating an attention mechanism for node aggregation specifically optimized for point cloud processing. The proposed GraphPT model surpasses the traditional truncated mean method in PID performance. In particular, for the CEPC baseline TPC, the <i>K/π</i> separation power improves by approximately 10% to 20% in the momentum interval from 5 to 20 GeV/<i>c</i>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

dN/dx reconstruction with deep learning for high-granularity TPCs

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月2日

查看全文
  • Guang Zhao,
  • Yue Chang,
  • Jinxian Zhang,
  • Linghui Wu,
  • Huirong Qi,
  • Xin She,
  • Mingyi Dong,
  • Shengsen Sun,
  • Jianchun Wang,
  • Yifang Wang,
  • Chunxu Yu

摘要

Particle identification (PID) is essential for future particle physics experiments such as the Circular Electron-Positron Collider (CEPC) and the Future Circular Collider. A high-granularity Time Projection Chamber (TPC) not only provides precise tracking but also enables dN/dx measurements for PID. The dN/dx method estimates the number of primary ionization electrons, offering significant improvements in PID performance. However, accurate reconstruction remains a major challenge for this approach. In this paper, we introduce a deep learning model, the Graph Point Transformer (GraphPT), for dN/dx reconstruction. In our approach, TPC data are represented as point clouds. The network backbone adopts a U-Net architecture built upon graph neural networks, incorporating an attention mechanism for node aggregation specifically optimized for point cloud processing. The proposed GraphPT model surpasses the traditional truncated mean method in PID performance. In particular, for the CEPC baseline TPC, the K/π separation power improves by approximately 10% to 20% in the momentum interval from 5 to 20 GeV/c.