TY - GEN
T1 - Discriminative Feature Learning for Video Semantic Segmentation
AU - Zhang, Han
AU - Jiang, Kai
AU - Zhang, Yu
AU - Li, Qing
AU - Xia, Changqun
AU - Chen, Xiaowu
N1 - Publisher Copyright:
© 2014 IEEE.
PY - 2015/9/28
Y1 - 2015/9/28
N2 - In this paper, we propose a novel deep learning based method for video semantic segmentation. Specially, we utilize 3D convolution neural network (3D CNN) to learn discriminative hierarchical features from spatial-temporal volumes for accurate pixel labelling. The learned features are capable of capturing both appearance and motion information. To align the pixel labels along real object boundaries, as well as maintain local consistency, we further perform graph-cut on a graph constructed on coherent 3d regions, or super-voxels, extracted from input video. Experiments demonstrate that due to the discriminative capability of learned features, our approach can obtain competitive labelling accuracy compared to the state-of-art in absence of sophisticated inference models, even with few training samples.
AB - In this paper, we propose a novel deep learning based method for video semantic segmentation. Specially, we utilize 3D convolution neural network (3D CNN) to learn discriminative hierarchical features from spatial-temporal volumes for accurate pixel labelling. The learned features are capable of capturing both appearance and motion information. To align the pixel labels along real object boundaries, as well as maintain local consistency, we further perform graph-cut on a graph constructed on coherent 3d regions, or super-voxels, extracted from input video. Experiments demonstrate that due to the discriminative capability of learned features, our approach can obtain competitive labelling accuracy compared to the state-of-art in absence of sophisticated inference models, even with few training samples.
KW - Convolutional neural network
KW - Deep learning
KW - Super-voxel
KW - Video semantic segmentation
UR - https://www.scopus.com/pages/publications/84962201551
U2 - 10.1109/ICVRV.2014.65
DO - 10.1109/ICVRV.2014.65
M3 - 会议稿件
AN - SCOPUS:84962201551
T3 - Proceedings - 2014 International Conference on Virtual Reality and Visualization, ICVRV 2014
SP - 321
EP - 326
BT - Proceedings - 2014 International Conference on Virtual Reality and Visualization, ICVRV 2014
A2 - Shen, Xukun
A2 - Zhang, Xiaopeng
A2 - Zhou, Zhong
A2 - Zhang, Guodong
A2 - Luo, Xun
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - International Conference on Virtual Reality and Visualization, ICVRV 2014
Y2 - 30 August 2014 through 31 August 2014
ER -