全文总字数:2733字
1. 毕业设计(论文)主要内容:
1.前期知识储备: 通过阅读相关文献,了解BSN(Boundary-Sensitive Network)模型、时序动作提议原理、python、PaddlePaddle、动作检测与分类等相关知识。
2.设计功能:本设计的任务包括: 1)基于视频数据(主要是室内监控视频),对其进行特征提取; 2)对于提取的视频特征,生成初步的时序动作提名结果; 3)对结果进行非极大化抑制,从而去除重叠的部分,得到最终的提名结果。
4)对于1)2)3),在PaddlePaddle环境中实现所设计和改进的BSN算法并对结果进行分析。
2. 毕业设计(论文)主要任务及要求
1.查阅15篇相关文献(不少于5篇外文文献),并每篇书写200—300字文献摘要(装订成册,带封面); 2.认真填写周记,完成至少1500字开题报告(“设计的目的及意义”至少800汉字;“基本内容和技术方案”至少400汉字;进度安排应尽可能详细;); 3.完成5000中文字以上的相关英文专业文献翻译,并装订成册(中英文一起,带封面); 4.完成方法研究、算法设计与实现; 5.按武汉理工大学理工类本科生毕业论文撰写规范撰写毕业论文,完成10000字以上的毕业论文; 6.进行论文答辩。
3. 毕业设计(论文)完成任务的计划与安排
1.2020/1/11—2020/1/24:明确选题,查阅相关文献,外文翻译和撰写开题报告; 2.2020/1/25—2020/4/30:系统架构,系统设计与开发(或算法研究与设计)、系统测试、分析、比较与完善; 3.2020/5/1—2020/5/25:撰写论文初稿;修改论文,定稿并提交论文评审; 4.2020/5/26—2020/6/6:准备论文答辩。
4. 主要参考文献
[1] Tianwei Lin, Xu Zhao, Haisheng Su, Chongjing Wang, Ming Yang. BSN: Boundary Sensitive Network for Temporal Action Proposal Generation[C]. Piscataway: ECCV,2018.[2] Tianwei Lin, Xiao Liu, Xin Li, Errui Ding, Shilei Wen. BMN: Boundary-Matching Network for Temporal Action Proposal Generation[C]. Piscataway: ICCV,2019.[3] Yue Zhao, Yuanjun Xiong, Limin Wang, Zhirong Wu, Xiaoou Tang, Dahua Lin. Temporal Action Detection with Structured Segment Networks[C]. Piscataway: ICCV,2019.[4] Yuanjun Xiong, Yue Zhao, Limin Wang, Dahua Lin, Xiaoou Tang .A Pursuit of Temporal Accuracy in General Activity Detection[C]. Piscataway: CVPR,2017.[5] Limin Wang, Yuanjun Xiong, Zhe Wang, Yu Qiao, Dahua Lin, Xiaoou Tang, Luc Van Gool .Temporal Segment Networks: Towards Good Practices for Deep Action Recognition[C]. Piscataway: ECCV,2016.[6] Zhenzhong Lan, Yi Zhu, Alexander G. Hauptmann. Deep Local Video Feature for Action Recognition[C]. Piscataway: CVPR,2017.[7] Bolei Zhou, Alex Andonian, Aude Oliva, Antonio Torralba.Temporal Relational Reasoning in Videos[C]. Piscataway: ECCV,2018.[8] Wangjiang Zhu, Jie Hu, Gang Sun, Xudong Cao, Yu Qiao. A Key Volume Mining Deep Framework for Action Recognition[C]. Piscataway: CVPR,2016.[9] Amlan Kar, Nishant Rai, Karan Sikka, Gaurav Sharma.AdaScan: Adaptive Scan Pooling in Deep Convolutional Neural Networks for Human Action Recognition in Videos[C]. Piscataway: CVPR,2017.[10] Chao Li, Qiaoyong Zhong, Di Xie, Shiliang Pu.Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation[C]. Piscataway: IJCAI,2018.[11] Sijie Yan, Yuanjun Xiong, Dahua Lin.Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition[C]. Piscataway: AAAI,2018.[12] Yansong Tang, Yi Tian, Jiwen Lu, Peiyang Li1 Jie Zhou.Deep Progressive Reinforcement Learning for Skeleton-based Action Recognition [C]. Piscataway: CVPR,2018.
以上是毕业论文任务书,课题毕业论文、开题报告、外文翻译、程序设计、图纸设计等资料可联系客服协助查找。