Kaifeng Su

Kaifeng Su

Auto-generated profile, based on publicly available information

Kaifeng Su

Northeastern University, Qinhuangdao, Hebei, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
28 September 2024 in Neural Networks
Verified
Journal Article
Closed access
CFChangzeng Fu
Changzeng Fu
FQFengkui Qian
Fengkui Qian
KSKaifeng Su
Kaifeng Su
YSYikai Su
Yikai Su
ZWZe Wang
Ze Wang
CICarlos Toshinori Ishi
Carlos Toshinori Ishi

12 December 2024 in Neurocomputing
Verified
Journal Article
Closed access
CFChangzeng Fu
Changzeng Fu
FQFengkui Qian
Fengkui Qian
YSYikai Su
Yikai Su
KSKaifeng Su
Kaifeng Su
SSSiyang Song
Siyang Song
HIHiroshi Ishiguro
Hiroshi Ishiguro

30 September 2024 in Expert Systems with Applications
Verified
Journal Article
Closed access
CFChangzeng Fu
Changzeng Fu
YSYikai Su
Yikai Su
KSKaifeng Su
Kaifeng Su
YLYinghao Liu
Yinghao Liu
JSJiaqi Shi
Jiaqi Shi
Hiroshi IshiguroHiroshi Ishiguro
Hiroshi Ishiguro

27 October 2025
Verified
Conference Paper
Open access
CFChangzeng Fu
Changzeng Fu
ZFZelin Fu
Zelin Fu
QZQi Zhang
Qi Zhang
XKXinhe Kuang
Xinhe Kuang
JDJiacheng Dong
Jiacheng Dong
Hiroshi IshiguroHiroshi Ishiguro
Hiroshi Ishiguro

21 July 2025 in Sensors
Verified
Journal Article
Open access
PDF file
SZShiwen Zhao
Shiwen Zhao
YZYunze Zhang
Yunze Zhang
YSYikai Su
Yikai Su
KSKaifeng Su
Kaifeng Su
JLJiemin Liu
Jiemin Liu
SYShiqi Yu
Shiqi Yu
The global prevalence of depression necessitates the application of technological solutions, particularly sensor-based systems, to augment scarce resources for early diagnostic purposes. In this study, we use benchmark datasets that contain multimodal data including video, audio, and transcribed text. To address depression detection as a chronic long-term disorder reflected by temporal behavioral patterns, we propose a novel framework that segments videos into utterance-level instances using GRU for contextual representation, and then constructs graphs where utterance embeddings serve as nodes connected through dual relationships capturing both chronological development and intermittent relevant information. Graph neural networks are employed to learn multi-dimensional edge relationships and align multimodal representations across different temporal dependencies. Our approach achieves superior performance with an MAE of 5.25 and RMSE of 6.75 on AVEC2014, and CCC of 0.554 and RMSE of 4.61 on AVEC2019, demonstrating significant improvements over existing methods that focus primarily on momentary expressions.