Chaolei Han

Chaolei Han (韩朝磊)
School of Cyber Science and Engineering, Southeast University
Address: 2 Southeast Road, Jiangning District, Nanjing, China
Email: chaoleihan98@gmail.com

About Me [GitHub] [Google Scholar]

I am currently pursuing the Ph.D. degree of Southeast University(SEU), Nanjing, China. Before that, I received M.S. degree in Nanjing Normal University(NNU).

My research interests include:

human activity recognition
computer vision
deep learning
machine learning

Honor

2022 Graduate National Scholarship
2023 Provincial Merit Student
2023 President Scholarship
The 11th APMCM Asia-Pacific Student Mathematical Modeling Competition, First Prize
The 7th Mathematical Modeling Challenge for College Students, First Prize
MathorCup Efficient Mathematical Modeling Challenge, Second Prize
The 3rd National Student Algorithm Design and Programming Challenge, Bronze medal
China Power Industry Campus Elite Challenge, Third Prize
The 4th "Transfusion Cup" National Student IT Skills Competition, Third Prize
2021 National Student English Translation Competition, Provincial Third Prize
The 17th "Challenge Cup" "B&B Technology" Special Competition, Provincial Third Prize

News

2023-04: 1 paper (Discrete Cosine Transform Convolution) accepted in IEEE Transactions on Knowledge and Data Engineering.
2023-04: 1 paper (Collaborative Compression Scheme) accepted in IEEE Transactions on Mobile Computing.
2023-03: 1 paper (Diverse-Branch Convolution) accepted in IEEE Transactions on Instrumentation and Measurement.
2022-07: 1 paper (Temporal-Aware and Modality-Aware Attention) accepted in IEEE Transactions on Instrumentation and Measurement.
2022-02: 1 paper (Heterogeneous Convolution) accepted in Expert Systems with Applications.

Selected Publications

(* indicates equal contribution, # corresponding author)

Human Activity Recognition Using Wearable Sensors by Heterogeneous Convolutional Neural Networks
Chaolei Han, Zhang Lei#, Yin Tang, Wenbo Huang, Fuhong Min, Jun He
in ESWA, 2022
[Paper] [BibTex] [Code]

The heterogeneous convolutions can complement each other, which is very beneficial to augment the receptive field of sensor signals for HAR application.

Understanding and Improving Channel Attention for Human Activity Recognition by Temporal-aware and Modality-aware Embedding
Chaolei Han, Zhang Lei#, Yin Tang, Shige Xu, Fuhong Min, Hao Wu, Aiguo Song
in TIM, 2022
[Paper] [BibTex]
TAMA might better interpret deep models' behavior by simultaneously highlighting the more distinguishable temporal intervals and sensor modalities to specific activities.