摘 要
在智能视频监控系统的发展过程中，行人身份识别成为当前研究的热点之一。基于深度学习的行人重识别技术因其在特征提取和模型训练方面的优势而受到研究者的关注。文章首先介绍了行人重识别的定义、应用领域以及基于深度学习的行人重识别技术的发展历程和现状。然后，本文概述了基于深度学习的行人重识别方法和常见的任务及解决方法。接着，讨论了当前基于深度学习的行人重识别技术改进研究的主要方向，包括模型优化与训练技巧的改进、特征提取与降维技术的改进以及跨数据集域适配与模型迁移技术的改进。最后，文章探讨了基于深度学习的行人重识别未来发展趋势，包括基于多模态信息和多任务学习的行人重识别方法、基于知识图谱和领域迁移学习的行人重识别方法以及基于生成式对抗网络和弱监督学习的行人重识别方法等。
关键词：智能 、视频监控 、行人、重识别
Abstract
With the continuous development of intelligent video surveillance technology, pedestrian recognition, as one of the important directions, has been increasingly widely applied. The pedestrian recognition technology based on deep learning has attracted researchers' attention due to its advantages in feature extraction and model training. The article first introduces the definition and application fields of pedestrian re recognition, as well as the development process and current status of pedestrian re recognition technology based on deep learning. Then, this article provides an overview of pedestrian re recognition methods based on deep learning and common tasks and solutions. Next, the main directions of current research on deep learning based pedestrian recognition technology improvement were discussed, including model optimization and training techniques improvement, feature extraction and dimensionality reduction techniques improvement, and cross dataset domain adaptation and model transfer techniques improvement. Finally, the paper discusses the future development trend of pedestrian re identification based on deep learning, including the pedestrian re identification method based on multimodal information and multi task learning, the pedestrian re identification method based on knowledge atlas and domain transfer learning, and the pedestrian re identification method based on generative confrontation network and weak supervised learning.
Keyword:Intelligent、video surveillance、pedestrian、re identification
目 录
引言 1
1深度学习在行人重识别相关概述 1
1.1行人重识别的定义 1
1.2基于深度学习的行人重识别技术的发展历程和现状 1
2基于深度学习的行人重识别方法综述 2
2.1深度学习技术的基本原理和分类 2
2.2基于深度学习的行人重识别方法及其应用 2
2.3常见的行人重识别任务及其解决方法 2
3基于深度学习的行人重识别技术改进研究 3
3.1模型优化与训练技巧的改进研究 3
3.2特征提取与降维技术的改进研究 3
3.3跨数据集域适配与模型迁移技术的改进研究 3
4基于深度学习的行人重识别未来发展趋势 4
4.1基于多模态信息和多任务学习的行人重识别方法 4
4.2基于知识图谱和领域迁移学习的行人重识别方法 4
4.3基于生成式对抗网络和弱监督学习的行人重识别方法 4
结论 5
参考文献 6
致谢 7
在智能视频监控系统的发展过程中，行人身份识别成为当前研究的热点之一。基于深度学习的行人重识别技术因其在特征提取和模型训练方面的优势而受到研究者的关注。文章首先介绍了行人重识别的定义、应用领域以及基于深度学习的行人重识别技术的发展历程和现状。然后，本文概述了基于深度学习的行人重识别方法和常见的任务及解决方法。接着，讨论了当前基于深度学习的行人重识别技术改进研究的主要方向，包括模型优化与训练技巧的改进、特征提取与降维技术的改进以及跨数据集域适配与模型迁移技术的改进。最后，文章探讨了基于深度学习的行人重识别未来发展趋势，包括基于多模态信息和多任务学习的行人重识别方法、基于知识图谱和领域迁移学习的行人重识别方法以及基于生成式对抗网络和弱监督学习的行人重识别方法等。
关键词：智能 、视频监控 、行人、重识别
Abstract
With the continuous development of intelligent video surveillance technology, pedestrian recognition, as one of the important directions, has been increasingly widely applied. The pedestrian recognition technology based on deep learning has attracted researchers' attention due to its advantages in feature extraction and model training. The article first introduces the definition and application fields of pedestrian re recognition, as well as the development process and current status of pedestrian re recognition technology based on deep learning. Then, this article provides an overview of pedestrian re recognition methods based on deep learning and common tasks and solutions. Next, the main directions of current research on deep learning based pedestrian recognition technology improvement were discussed, including model optimization and training techniques improvement, feature extraction and dimensionality reduction techniques improvement, and cross dataset domain adaptation and model transfer techniques improvement. Finally, the paper discusses the future development trend of pedestrian re identification based on deep learning, including the pedestrian re identification method based on multimodal information and multi task learning, the pedestrian re identification method based on knowledge atlas and domain transfer learning, and the pedestrian re identification method based on generative confrontation network and weak supervised learning.
Keyword:Intelligent、video surveillance、pedestrian、re identification
目 录
引言 1
1深度学习在行人重识别相关概述 1
1.1行人重识别的定义 1
1.2基于深度学习的行人重识别技术的发展历程和现状 1
2基于深度学习的行人重识别方法综述 2
2.1深度学习技术的基本原理和分类 2
2.2基于深度学习的行人重识别方法及其应用 2
2.3常见的行人重识别任务及其解决方法 2
3基于深度学习的行人重识别技术改进研究 3
3.1模型优化与训练技巧的改进研究 3
3.2特征提取与降维技术的改进研究 3
3.3跨数据集域适配与模型迁移技术的改进研究 3
4基于深度学习的行人重识别未来发展趋势 4
4.1基于多模态信息和多任务学习的行人重识别方法 4
4.2基于知识图谱和领域迁移学习的行人重识别方法 4
4.3基于生成式对抗网络和弱监督学习的行人重识别方法 4
结论 5
参考文献 6
致谢 7
