Mei Chee Leong
Mei Chee Leong has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Dilip Kumar Prasad
UiT The Arctic University of Norway, Tromsø, Norway
2 shared publications - Feng Lin
China-Singapore International Joint Research Institute
2 shared publications - Hui Li Tan
Institute for Infocomm Research (I2R), Agency for Science, Technology, and Research (A*STAR), Singapore, Singapore
2 shared publications - Liyuan Li
Institute for Infocomm Research
2 shared publications - Nancy F. Chen
Agency for Science, Technology and Research (A*STAR), Singapore, Singapore
2 shared publications - Ying Gu
Institute of Advanced Intelligence and Computing (IAIC)
2 shared publications
Latest Publications
- 1 October 2026
- Published by Springer Nature
- p. 298-318
Explicit Logic Channel for Validation and Enhancement of MLLMs on Zero-Shot Tasks
- 1 October 2026
- Published by Springer Nature
- p. 298-318
Explicit Logic Channel for Validation and Enhancement of MLLMs on Zero-Shot Tasks
- Open Access
- 12 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 185-189
Contextualized Visual Storytelling for Conversational Chatbot in Education
- 12 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 185-189
Open AccessContextualized Visual Storytelling for Conversational Chatbot in Education
- 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.
- Open Access
- 7 December 2019
- Published by MDPI AG in Preprints.org
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 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.
- 7 December 2019
- Published by MDPI AG in Preprints.org
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 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.