Zhang Fan, Théberge Antoine, Jodoin Pierre-Marc, Descoteaux Maxime, O'Donnell Lauren J
University of Electronic Science and Technology of China, Chengdu, China.
Sherbrooke Connectivity Imaging Laboratory (SCIL), Université de Sherbrooke, Sherbrooke, QC, Canada.
Brain Struct Funct. 2025 Jun 16;230(6):100. doi: 10.1007/s00429-025-02938-0.
Tractography is a challenging process with complex rules, driving continuous algorithmic evolution to address its challenges. Meanwhile, deep learning has tackled similarly difficult tasks, such as mastering the Go board game and animating sophisticated robots. Given its transformative impact in these areas, deep learning has the potential to revolutionize tractography within the framework of existing rules. This work provides a brief summary of recent advances and challenges in deep learning-based tractography computing and analysis.
纤维束成像技术是一个具有复杂规则的挑战性过程，推动着算法不断演进以应对其挑战。与此同时，深度学习已经攻克了类似的难题，比如掌握围棋游戏和操控复杂的机器人。鉴于其在这些领域的变革性影响，深度学习有潜力在现有规则框架内彻底改变纤维束成像技术。本文简要总结了基于深度学习的纤维束成像计算与分析的最新进展和挑战。
