RRNet: Repetition-Reduction Network for Energy Efficient Decoder of Depth Estimation
arXiv:1907.09707
Abstract
We introduce Repetition-Reduction network (RRNet) for resource-constrained depth estimation, offering significantly improved efficiency in terms of computation, memory and energy consumption. The proposed method is based on repetition-reduction (RR) blocks. The RR blocks consist of the set of repeated convolutions and the residual connection layer that take place of the pointwise reduction layer with linear connection to the decoder. The RRNet help reduce memory usage and power consumption in the residual connections to the decoder layers. RRNet consumes approximately 3.84 times less energy and 3.06 times less meory and is approaximately 2.21 times faster, without increasing the demand on hardware resource relative to the baseline network (Godard et al, CVPR'17), outperforming current state-of-the-art lightweight architectures such as SqueezeNet, ShuffleNet, MobileNet and PyDNet.
arXiv comment9 pages, 5 figures
References in corpus (13)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
- Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
- Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
- DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
- Xception: Deep Learning with Depthwise Separable Convolutions
- Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue
- Compression of Deep Convolutional Neural Networks for Fast and Low Power Mobile Applications
- Fast Algorithms for Convolutional Neural Networks
- Pyramid Stereo Matching Network
- Efficient Sparse-Winograd Convolutional Neural Networks
- Learning for Disparity Estimation through Feature Constancy