Dual Skipping Networks
Abstract
Inspired by the recent neuroscience studies on the left-right asymmetry of the human brain in processing low and high spatial frequency information, this paper introduces a dual skipping network which carries out coarse-to-fine object categorization. Such a network has two branches to simultaneously deal with both coarse and fine-grained classification tasks. Specifically, we propose a layer-skipping mechanism that learns a gating network to predict which layers to skip in the testing stage. This layer-skipping mechanism endows the network with good flexibility and capability in practice. Evaluations are conducted on several widely used coarse-to-fine object categorization benchmarks, and promising results are achieved by our proposed network model.
Cite
@article{arxiv.1710.10386,
title = {Dual Skipping Networks},
author = {Changmao Cheng and Yanwei Fu and Yu-Gang Jiang and Wei Liu and Wenlian Lu and Jianfeng Feng and Xiangyang Xue},
journal= {arXiv preprint arXiv:1710.10386},
year = {2018}
}
Comments
CVPR 2018 (poster); fix typo