ESDAR-net: towards high-accuracy and real-time driver action recognition for embedded systems.
Yaocong HuZhen ShuaiHuicheng YangGuoyang WanYajun ZhangChao XieMingQi LuXiaobo LuPublished in: Multim. Tools Appl. (2024)
Keyphrases
- action recognition
- embedded systems
- real time
- real time systems
- low cost
- high accuracy
- resource limited
- human actions
- computing power
- depth cameras
- activity recognition
- bag of words
- safety critical
- embedded devices
- human detection
- action classification
- embedded software
- computer vision
- recognition of human actions
- embedded real time systems
- recognizing human actions
- software systems
- static images
- human pose
- body parts
- cyber physical systems
- depth sensors
- action primitives
- recognizing actions
- quality of service
- view invariant
- video dataset
- field programmable gate array
- human activities
- bag of features
- multi agent systems
- view invariant action recognition