基于3D密集连接的异常行为识别
Wei Chen1, Zhanhe Yu2, Chaochao Yang1
1School of Electrical and Control Engineering, North China University of Technology, Beijing 100144, P. R. China.
International journal of neural systems
|July 16, 2024
概括
本研究介绍了一种新的异常行为识别算法,使用3D密集连接进行增强的实时检测. 该方法通过快速识别与正常模式的偏差,以高的准确性和效率来提高城市安全.
科学领域:
- 计算机科学 计算机科学
- 人工智能的人工智能
- 机器学习 机器学习
背景情况:
- 识别异常行为对于安全和欺诈检测至关重要.
- 深度卷积网络 (ConvNets) 是有前途的,但缺乏实时功能.
- 现有的方法优先考虑准确性而不是速度,阻碍了立即威胁的识别.
研究的目的:
- 开发一个实时异常行为识别算法.
- 通过快速检测异常活动,增强城市公共安全.
- 通过解决计算效率来改进现有的基于ConvNet的方法.
主要方法:
- 提出了一个基于三维 (3D) 密集连接的算法.
- 利用多实例学习策略来分类各种异常行为.
- 采用密集连接模块和软值注意力机制来优化模型参数和计算效率.
- 实施了注意分配,以减少冗余的顺序信息.
主要成果:
- 在UCF犯罪数据集上实现了95.61%的识别准确度.
- 与现有模型相比,在识别精度和处理速度方面表现出强的性能.
- 成功降低了模型参数数量,提高了网络计算效率.
结论:
- 提出的3D密集连接算法为实时异常行为识别提供了有效的解决方案.
- 该方法平衡了高精度与提高计算效率,这对于公共安全应用至关重要.
- 注意力机制和密集的连接有助于减轻冗余数据对性能的影响.
相关概念视频
01:25Space Trusses
A space truss is a three-dimensional counterpart of a planar truss. These structures consist of members connected at their ends, often utilizing ball-and-socket joints to create a stable and versatile framework. The space truss is widely used in various construction projects due to its adaptability and capacity to withstand complex loads.
At the core of a space truss lies the fundamental unit known as the tetrahedron. This structure is composed of six members that form a three-dimensional shape...
At the core of a space truss lies the fundamental unit known as the tetrahedron. This structure is composed of six members that form a three-dimensional shape...
01:29Three-Dimensional Analysis of Strain
Three-dimensional strain analysis is crucial for understanding how materials deform under stress, particularly in elastic, homogeneous materials. This method employs principal stress axes to simplify complex stress states into more understandable forms. Subjected to stress, a small cubic element within a material either expands or contracts along these axes, transforming into a rectangular parallelepiped. This transformation effectively illustrates the material's deformation. The principal...