Fixed-Point Quantization of 3D Convolutional Neural Networks for Energy-Efficient Action Recognition.
Hyunhoon LeeYounghoon ByunSeokha HwangSunggu LeeYoungjoo LeePublished in: ISOCC (2018)
Keyphrases
- fixed point
- energy efficient
- action recognition
- convolutional neural networks
- wireless sensor networks
- energy consumption
- human actions
- sensor networks
- action classification
- activity recognition
- computer vision
- energy efficiency
- sufficient conditions
- recognition of human actions
- dynamical systems
- routing protocol
- base station
- human activities
- belief propagation
- recognizing human actions
- object recognition
- action recognition in videos
- routing algorithm
- machine learning
- reinforcement learning
- image processing