HNPU-V2: A 46.6 FPS DNN Training Processor for Real-World Environmental Adaptation based Robust Object Detection on Mobile Devices.
Donghyeon HanDongseok ImGwangtae ParkYoungwoo KimSeokchan SongJuhyoung LeeHoi-Jun YooPublished in: HCS (2022)
Keyphrases
- object detection
- mobile devices
- real world
- training process
- high speed
- wide range
- data sets
- discriminatively trained
- mobile phone
- online learning
- real time
- context aware
- mobile computing
- frame rate
- mobile learning
- face detection
- parallel processing
- partial occlusion
- supervised learning
- object detectors
- covariate shift
- training samples
- background subtraction
- mobile applications
- mobile users
- mobile technologies
- mobile environments