A 141.4 mW Low-Power Online Deep Neural Network Training Processor for Real-time Object Tracking in Mobile Devices.
Donghyeon HanJinsu LeeJinmook LeeSungpill ChoiHoi-Jun YooPublished in: ISCAS (2018)
Keyphrases
- low power
- power consumption
- neural network training
- single chip
- high speed
- mobile devices
- real time object tracking
- gate array
- low cost
- neural network
- logic circuits
- training algorithm
- low power consumption
- power reduction
- parallel processing
- power dissipation
- genetic algorithm
- real time
- image sensor
- particle swarm optimisation
- computer vision
- mixed signal