Lin Dong

Lin Dong

Auto-generated profile, based on publicly available information

Lin Dong

Westlake Institute for Data Intelligence, Hangzhou, 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
5 May 2022 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
HDHuanzhang Dou
Huanzhang Dou
PZPengyi Zhang
Pengyi Zhang
YZYuhan Zhao
Yuhan Zhao
LDLin Dong
Lin Dong
ZQZequn Qin
Zequn Qin
XLXi Li
Xi Li
Gait recognition aims at identifying the pedestrians at a long distance by their biometric gait patterns. It is inherently challenging due to the various covariates and the properties of silhouettes (textureless and colorless), which result in two kinds of pair-wise hard samples: the same pedestrian could have distinct silhouettes (intra-class diversity) and different pedestrians could have similar silhouettes (inter-class similarity). In this work, we propose to solve the hard sample issue with a Memory-augmented Progressive Learning network (GaitMPL), including Dynamic Reweighting Progressive Learning module (DRPL) and Global Structure-Aligned Memory bank (GSAM). Specifically, DRPL reduces the learning difficulty of hard samples by easy-to-hard progressive learning. GSAM further augments DRPL with a structure-aligned memory mechanism, which maintains and models the feature distribution of each ID. Experiments on two commonly used datasets, CASIA-B and OU-MVLP, demonstrate the effectiveness of GaitMPL. On CASIA-B, we achieve the state-of-the-art performance, i.e., 88.0% on the most challenging condition (Clothing) and 93.3% on the average condition, which outperforms the other methods by at least 3.8% and 1.4%, respectively.

28 January 2023 in npj Urban Sustainability
Verified
Journal Article
Open access
PDF file
HZHongshen Zhang
Hongshen Zhang
YZYongtao Zhang
Yongtao Zhang
Shibo HeShibo He
Shibo He
YFYi Fang
Yi Fang
YCYanggang Cheng
Yanggang Cheng
JCJiming Chen
Jiming Chen
Currently, the global situation of COVID-19 is aggravating, pressingly calling for efficient control and prevention measures. Understanding the spreading pattern of COVID-19 has been widely recognized as a vital step for implementing non-pharmaceutical measures. Previous studies explained the differences in contagion rates due to the urban socio-political measures, while fine-grained geographic urban spreading pattern still remains an open issue. Here, we fill this gap by leveraging the trajectory data of 197,808 smartphone users (including 17,808 anonymous confirmed cases) in nine cities in China. We find a general spreading pattern in all cities: the spatial distribution of confirmed cases follows a power-law-like model and the spreading centroid human mobility is time-invariant. Moreover, we reveal that long average traveling distance results in a high growth rate of spreading radius and wide spatial diffusion of confirmed cases in the fine-grained geographic model. With such insight, we adopt the Kendall model to simulate the urban spreading of COVID-19 which can well fit the real spreading process. Our results unveil the underlying mechanism behind the spatial-temporal urban evolution of COVID-19, and can be used to evaluate the performance of mobility restriction policies implemented by many governments and to estimate the evolving spreading situation of COVID-19.

25 April 2022 in Journal of Seismology
Verified
Journal Article
Closed access
PDF file
ZZZengwei Zheng
Zengwei Zheng
JWJiquan Wang
Jiquan Wang
LSLifei Shi
Lifei Shi
SZSha Zhao
Sha Zhao
JHJianmin Hou
Jianmin Hou
LDLin Dong
Lin Dong

6 June 2023 in arXiv
Verified
Preprint
Open access
PDF file
HDHuanzhang Dou
Huanzhang Dou
PZPengyi Zhang
Pengyi Zhang
YZYuhan Zhao
Yuhan Zhao
LDLin Dong
Lin Dong
ZQZequn Qin
Zequn Qin
XLXi Li
Xi Li
Gait recognition aims at identifying the pedestrians at a long distance by their biometric gait patterns. It is inherently challenging due to the various covariates and the properties of silhouettes (textureless and colorless), which result in two kinds of pair-wise hard samples: the same pedestrian could have distinct silhouettes (intra-class diversity) and different pedestrians could have similar silhouettes (inter-class similarity). In this work, we propose to solve the hard sample issue with a Memory-augmented Progressive Learning network (GaitMPL), including Dynamic Reweighting Progressive Learning module (DRPL) and Global Structure-Aligned Memory bank (GSAM). Specifically, DRPL reduces the learning difficulty of hard samples by easy-to-hard progressive learning. GSAM further augments DRPL with a structure-aligned memory mechanism, which maintains and models the feature distribution of each ID. Experiments on two commonly used datasets, CASIA-B and OU-MVLP, demonstrate the effectiveness of GaitMPL. On CASIA-B, we achieve the state-of-the-art performance, i.e., 88.0% on the most challenging condition (Clothing) and 93.3% on the average condition, which outperforms the other methods by at least 3.8% and 1.4%, respectively.