EIREducational Innovation and Research3066-82983066-828XArt and Technology10.61369/EIR.2026060032Article基于Agent Skill的高校课程项目选题研究—— 以《数字图像处理与计算机视觉》为例https://artdesignp.com/journal/EIR/2/6/10.61369/EIR.2026060032李智慧,刘咏梅2026262026-06-20针对《数字图像处理与计算机视觉》课程中学生选题困难、易选现成项目的问题，本文提出基于Agent Skill的自动化选题Topic-select Skill。该Skill通过双策略方式选题，策略A通过检索小众开源论文、分析论文中的缺陷从而发现改进空间；策略B通过搜索SOTA方法的应用空白领域、评估跨域迁移可行性以生成创新选题。系统对候选项目进行难度可行性、代码量、可改进空间、可复现性四维度综合评分，最后输出排序的推荐列表。实验以5道选题为验证案例，验证了方法的有效性与实用性。Agent Skill,课程项目选题,计算机视觉,深度学习教学

[1] 张剑平, 陈仕品. 人工智能赋能教育变革：机遇、挑战与路径[J]. 中国电化教育, 2020, （1）: 1-8.[2] 余胜泉. 人工智能教师的未来角色[J]. 开放教育研究, 2018, 24（1）: 16-28.[3] Kasneci E, et al. ChatGPT for good? On opportunities and challenges of large language models for education[J]. Learning and Individual Differences, 2023, 103: 102274.[4] Chu Z, Wang S, Xie J, et al. LLM agents for education: Advances and applications[J]. arXiv preprint arXiv:2503.11733, 2025, 2.[5] 黄荣怀, 周伟, 杜静, 等. 面向智能教育的三个基本计算问题[J]. 开放教育研究, 2019, 25（5）: 11-22.[6] Chen P, et al. KnowEdu: A system to construct knowledge graph for education[J]. IEEE Access, 2018, 6: 31553-31563.[7] Ghosh A, et al. The future of assessments in the age of generative AI[J]. Journal of Computer Assisted Learning, 2024, 40（3）: 1012-1025.[8] Leinonen J, et al. Comparing code explanations created by students and large language models[C]//ITiCSE 2023: 124-130.[9] Papers with Code. Machine learning papers with code, evaluation tables and benchmarks[EB/OL]. https://paperswithcode.com/, 2024.[10] Zhongwei V. Hermes Edu Skills: Agent Skills for Chinese education scenarios[EB/OL]. https://github.com/zhongweiv/hermes-edu-Skills, 2025.