Yong Tsui Lee
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yong Tsui Lee has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
Fen Fang
5 shared publications- Feng Lin
China-Singapore International Joint Research Institute
2 shared publications - H. L. Zou
School of Mechanical and Aerospace Engineering, Nanyang Technological University, Nanyang Avenue, Singapore 639798, Singapore
2 shared publications - Dilip Kumar Prasad
UiT The Arctic University of Norway, Tromsø, Norway
1 shared publication - Hock Soon Seah
Nanyang Technological University
1 shared publication - Jianping Cai
Nanyang Technological University
1 shared publication
Latest Publications
- Open Access
- 12 January 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (2) , 557
Semi-CNN Architecture for Effective Spatio-Temporal Learning in Action Recognition
This paper introduces a fusion convolutional architecture for efficient learning of spatio-temporal features in video action recognition. Unlike 2D convolutional neural networks (CNNs), 3D CNNs can be applied directly on consecutive frames to extract spatio-temporal features. The aim of this work is to fuse the convolution layers from 2D and 3D CNNs to allow temporal encoding with fewer parameters than 3D CNNs. We adopt transfer learning from pre-trained 2D CNNs for spatial extraction, followed by temporal encoding, before connecting to 3D convolution layers at the top of the architecture. We construct our fusion architecture, semi-CNN, based on three popular models: VGG-16, ResNets and DenseNets, and compare the performance with their corresponding 3D models. Our empirical results evaluated on the action recognition dataset UCF-101 demonstrate that our fusion of 1D, 2D and 3D convolutions outperforms its 3D model of the same depth, with fewer parameters and reduces overfitting. Our semi-CNN architecture achieved an average of 16–30% boost in the top-1 accuracy when evaluated on an input video of 16 frames.
- 12 January 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (2) , 557
Open AccessSemi-CNN Architecture for Effective Spatio-Temporal Learning in Action Recognition
This paper introduces a fusion convolutional architecture for efficient learning of spatio-temporal features in video action recognition. Unlike 2D convolutional neural networks (CNNs), 3D CNNs can be applied directly on consecutive frames to extract spatio-temporal features. The aim of this work is to fuse the convolution layers from 2D and 3D CNNs to allow temporal encoding with fewer parameters than 3D CNNs. We adopt transfer learning from pre-trained 2D CNNs for spatial extraction, followed by temporal encoding, before connecting to 3D convolution layers at the top of the architecture. We construct our fusion architecture, semi-CNN, based on three popular models: VGG-16, ResNets and DenseNets, and compare the performance with their corresponding 3D models. Our empirical results evaluated on the action recognition dataset UCF-101 demonstrate that our fusion of 1D, 2D and 3D convolutions outperforms its 3D model of the same depth, with fewer parameters and reduces overfitting. Our semi-CNN architecture achieved an average of 16–30% boost in the top-1 accuracy when evaluated on an input video of 16 frames.
- 24 March 2016
- Published by Springer Nature in The Visual Computer
- Vol. 33 (10) , 1307-1318
Modeling and dynamics simulation for deformable objects of orthotropic materials
- 24 March 2016
- Published by Springer Nature in The Visual Computer
- Vol. 33 (10) , 1307-1318
Modeling and dynamics simulation for deformable objects of orthotropic materials
- 1 December 2015
- Published by Elsevier in Pattern Recognition
- Vol. 48 (12) , 3825-3842
Identification of faces in line drawings by edge decomposition
- 1 December 2015
- Published by Elsevier in Pattern Recognition
- Vol. 48 (12) , 3825-3842
Identification of faces in line drawings by edge decomposition
- 1 June 2014
- Published by Elsevier in Computer-Aided Design
- Vol. 51, 18-30
Efficient decomposition of line drawings of connected manifolds without face identification
- 1 June 2014
- Published by Elsevier in Computer-Aided Design
- Vol. 51, 18-30
Efficient decomposition of line drawings of connected manifolds without face identification
- 1 November 2012
- Published by Elsevier in Computer-Aided Design
- Vol. 44 (11) , 1090-1102
A new hybrid method for 3D object recovery from 2D drawings and its validation against the cubic corner method and the optimisation-based method
- 1 November 2012
- Published by Elsevier in Computer-Aided Design
- Vol. 44 (11) , 1090-1102
A new hybrid method for 3D object recovery from 2D drawings and its validation against the cubic corner method and the optimisation-based method
- 11 July 2012
- Published by Springer Nature in 3D Research
- Vol. 3 (3) , 3
Optimization-based freeform shape reconstruction
- 11 July 2012
- Published by Springer Nature in 3D Research
- Vol. 3 (3) , 3
Optimization-based freeform shape reconstruction
- 4 April 2012
- Published by Springer Nature in 3D Research
- Vol. 3 (2) , 1-8
3D reconstruction of polyhedral objects from single perspective projections using cubic corner
- 4 April 2012
- Published by Springer Nature in 3D Research
- Vol. 3 (2) , 1-8
3D reconstruction of polyhedral objects from single perspective projections using cubic corner
- 6 April 2011
- Published by Elsevier in Computer-Aided Design
- Vol. 43 (8) , 1025-1034
3D reconstruction of polyhedral objects from single parallel projections using cubic corner
- 6 April 2011
- Published by Elsevier in Computer-Aided Design
- Vol. 43 (8) , 1025-1034
3D reconstruction of polyhedral objects from single parallel projections using cubic corner
- 1 November 2007
- Published by Elsevier in Computer-Aided Design
- Vol. 39 (11) , 1025-1036
Constraint-based beautification and dimensioning of 3D polyhedral models reconstructed from 2D sketches
- 1 November 2007
- Published by Elsevier in Computer-Aided Design
- Vol. 39 (11) , 1025-1036
Constraint-based beautification and dimensioning of 3D polyhedral models reconstructed from 2D sketches
- 1 January 2005
- Published by U-turn Press LLC in Computer-Aided Design and Applications
- Vol. 2 (1-4) , 393-402
Detecting Minimum Over-constrained Subgraphs in 2D and 3D Based on Degree of Freedom Analysis
We present a flow-based method for finding minimum over constrained subgraphs in a geometric constraint graph. In 2D several different approaches have been implemented, while the 3D problem, in which the entities are points, lines and planes of a model, is much less investigated in the literature. We present a general algorithm for the problem both in 2D and 3D, show that the algorithm is correct and can be used to identify if there is an over constrained subgraph in a constraint graph.
- 1 January 2005
- Published by U-turn Press LLC in Computer-Aided Design and Applications
- Vol. 2 (1-4) , 393-402
Detecting Minimum Over-constrained Subgraphs in 2D and 3D Based on Degree of Freedom Analysis
We present a flow-based method for finding minimum over constrained subgraphs in a geometric constraint graph. In 2D several different approaches have been implemented, while the 3D problem, in which the entities are points, lines and planes of a model, is much less investigated in the literature. We present a general algorithm for the problem both in 2D and 3D, show that the algorithm is correct and can be used to identify if there is an over constrained subgraph in a constraint graph.