Meng Dai

Meng Dai

Auto-generated profile, based on publicly available information

Meng Dai

Faculty of Information Technology, Beijing University of Technology, Beijing, China

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
22 January 2024 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
YZYu Zhang
Yu Zhang
SZSun Zhonghua
Sun Zhonghua
MDMeng Dai
Meng Dai
JFJinchao Feng
Jinchao Feng
KJKebin Jia
Kebin Jia
As skeleton data becomes increasingly available, Graph Convolutional Networks (GCNs) are popularly adapted to extract the spatial and temporal features for skeleton-based action recognition. However, there are still limitations to be addressed in GCN-based methods. First, the multi-level semantic features fail to be connected, making fine-grained information loss as the network deepens. Second, the cross-scale spatiotempral features fail to be simultaneously considered and refined to focus on informative areas. These limitations lead to the challenge in distinguishing the confusing actions. To address these issues, we propose a cross-scale connection (CSC) structure and a spatiotemporal refinement focus (STRF) module. The CSC aims to bridge the gap between multi-level semantic features. The STRF module refines the cross-scale spatiotemporal features to focus on informative joints in each frame. Both are embedded into the standard GCNs to form the cross-scale spatiotemporal refinement network (CSR-Net). Our proposed CSR-Net explicitly models the cross-scale spatiotemporal information among multi-level semantic representations to boost the distinguishing capability for ambiguous actions. We conduct extensive experiments to demonstrate the effectiveness of our proposed method and it outperforms state-of-the-art methods on the NTU RGB+D 60, NTU-RGB+D 120 and NW-UCLA datasets.

31 December 2022 in Signal, Image and Video Processing
Verified
Journal Article
Closed access
PDF file
SZSun Zhonghua
Sun Zhonghua
TWTianyi Wang
Tianyi Wang
MDMeng Dai
Meng Dai

17 November 2022
Verified
Conference Paper
Open access
SZSun Zhonghua
Sun Zhonghua
MDMeng Dai
Meng Dai
ZYZiwen Yi
Ziwen Yi
TWTianyi Wang
Tianyi Wang
JFJinchao Feng
Jinchao Feng
KJKebin Jia
Kebin Jia
Effective feature learning is one of the prime components for human action recognition algorithm. Three-dimensional convolutional neural network (3D CNN) can directly extract spatio-temporal features, however it is insufficient to capture the most discriminative part of the action video. The redundant spatial regions within and between temporal frames would weak the descriptive ability of the 3D CNN model. To address this problem, we propose a lightweight spatio-temporal attention module (ST-AM), composed of spatial attention module (SAM) and temporal attention module (TAM). SAM and TAM can effectively encode the semantic spatial areas and suppress the redundant temporal frames to reduce misclassification. The proposed SAM and TAM have complementary effects and can be easily embedded into the existing 3D CNN action recognition model. Experiment on UCF-101 and HMDB-51 datasets shows that the ST-AM embedded model achieves impressive performance on action recognition task.

17 November 2023
Verified
Conference Paper
Closed access
YZYu Zhang
Yu Zhang
SZSun Zhonghua
Sun Zhonghua
MDMeng Dai
Meng Dai
JFJinchao Feng
Jinchao Feng
KJKebin Jia
Kebin Jia
Few-shot action recognition is the task of predicting unlabeled actions based on a limited number of labeled actions. Recently most of few-shot action recognition methods adopt GCNs to encode the skeleton clips and predict the label by calculating the distance in spatial-temporal feature space. However, coarse encoded features will lead to confusion between ambiguous actions with similar spatial or temporal information, which results in misclassification among these actions. To solve this problem, we propose a effective method called Spatial-Temporal Decoupling Matching (STDM) consisting of a spatial-temporal decoupling module and a spatial-temporal matching strategy to model the spatial and temporal information effectively. The spatial-temporal decoupling module maps confusing encoded features to spatial and temporal feature spaces respectively. The spatial-temporal matching strategy simultaneously recognizes the actions from the spatial and temporal perspective. We have conducted numerous experiments on NTU RGB+D 120 and Kinetics datasets. Our method achieves comparable results and surpasses most of the SOTA methods.