TKD: Temporal Knowledge Distillation for Active Perception
arXiv:1903.01522
Abstract
Deep neural networks based methods have been proved to achieve outstanding performance on object detection and classification tasks. Despite significant performance improvement, due to the deep structures, they still require prohibitive runtime to process images and maintain the highest possible performance for real-time applications. Observing the phenomenon that human vision system (HVS) relies heavily on the temporal dependencies among frames from the visual input to conduct recognition efficiently, we propose a novel framework dubbed as TKD: temporal knowledge distillation. This framework distills the temporal knowledge from a heavy neural networks based model over selected video frames (the perception of the moments) to a light-weight model. To enable the distillation, we put forward two novel procedures: 1) an Long-short Term Memory (LSTM) based key frame selection method; and 2) a novel teacher-bounded loss design. To validate, we conduct comprehensive empirical evaluations using different object detection methods over multiple datasets including Youtube-Objects and Hollywood scene dataset. Our results show consistent improvement in accuracy-speed trad-offs for object detection over the frames of the dynamic scene, compare to other modern object recognition methods.
References in corpus (12)
- Distilling the Knowledge in a Neural Network
- SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
- YOLOv3: An Incremental Improvement
- R-FCN: Object Detection via Region-based Fully Convolutional Networks
- Deep Domain Confusion: Maximizing for Domain Invariance
- Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
- FitNets: Hints for Thin Deep Nets
- DSSD : Deconvolutional Single Shot Detector
- Compressing Deep Convolutional Networks using Vector Quantization
- DeepDriving: Learning Affordance for Direct Perception in Autonomous Driving
- Convolutional Neural Networks: Ensemble Modeling, Fine-Tuning and Unsupervised Semantic Localization for Intraoperative CLE Images
- When Unsupervised Domain Adaptation Meets Tensor Representations