MultiPhys: Heterogeneous Fusion of Mamba and Transformer for Video-Based Multi-Task Physiological Measurement DOI Creative Commons

Chen Huo,

Pengbo Yin, Bo Fu

et al.

Sensors, Journal Year: 2024, Volume and Issue: 25(1), P. 100 - 100

Published: Dec. 27, 2024

Due to its non-contact characteristics, remote photoplethysmography (rPPG) has attracted widespread attention in recent years, and been widely applied for physiological measurements. However, most of the existing rPPG models are unable estimate multiple signals simultaneously, performance limited available multi-task is also restricted due their single-model architectures. To address above problems, this study proposes MultiPhys, adopting a heterogeneous network fusion approach development. Specifically, Convolutional Neural Network (CNN) used quickly extract local features early stage, transformer captures global context long-distance dependencies, Mamba compensate transformer’s deficiencies, reducing computational complexity improving accuracy model. Additionally, gate utilized feature selection, which classifies different indicators. Finally, indicators estimated after passing each task-related head. Experiments on three datasets show that MultiPhys superior handling tasks. The results cross-dataset hyper-parameter sensitivity tests verify generalization ability robustness, respectively. can be considered as an effective solution estimation, thus promoting development field.

Language: Английский

Quantifying the Impacts of Climate Change and Human Activities on Runoff in the Upper Yongding River Basin DOI
Yiyang Yang, Siyu Cai, Xiangyu Sun

et al.

Journal of Hydrologic Engineering, Journal Year: 2025, Volume and Issue: 30(2)

Published: Jan. 6, 2025

Language: Английский

Citations

0
Short-Term Water Level Prediction for Long-Distance Water Diversion Projects Using Data-Driven Methods with Multi-Scale Attention Mechanism DOI
Xinyong Xu,

Zhongkui Zhu,

Xiaonan Chen

et al.

Water Resources Management, Journal Year: 2025, Volume and Issue: unknown

Published: March 12, 2025

Language: Английский

Citations

0
Machine Learning Methods for the Prediction of Wastewater Treatment Efficiency and Anomaly Classification with Lack of Historical Data DOI Creative Commons
Igor Gulshin, Olga Kuzina

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(22), P. 10689 - 10689

Published: Nov. 19, 2024

This study examines an algorithm for collecting and analyzing data from wastewater treatment facilities, aimed at addressing regression tasks predicting the quality of treated classification preventing emergency situations, specifically filamentous bulking activated sludge. The feasibility using obtained under laboratory conditions simulating technological process as a training dataset is explored. A small collected actual plants considered test dataset. For both tasks, best results were achieved gradient-boosting models CatBoost family, yielding metrics SMAPE = 9.1 ROC-AUC 1.0. set most important predictors modeling was selected each target features.

Language: Английский

Citations

3
MultiPhys: Heterogeneous Fusion of Mamba and Transformer for Video-Based Multi-Task Physiological Measurement DOI Creative Commons

Chen Huo,

Pengbo Yin, Bo Fu

et al.

Sensors, Journal Year: 2024, Volume and Issue: 25(1), P. 100 - 100

Published: Dec. 27, 2024

Due to its non-contact characteristics, remote photoplethysmography (rPPG) has attracted widespread attention in recent years, and been widely applied for physiological measurements. However, most of the existing rPPG models are unable estimate multiple signals simultaneously, performance limited available multi-task is also restricted due their single-model architectures. To address above problems, this study proposes MultiPhys, adopting a heterogeneous network fusion approach development. Specifically, Convolutional Neural Network (CNN) used quickly extract local features early stage, transformer captures global context long-distance dependencies, Mamba compensate transformer’s deficiencies, reducing computational complexity improving accuracy model. Additionally, gate utilized feature selection, which classifies different indicators. Finally, indicators estimated after passing each task-related head. Experiments on three datasets show that MultiPhys superior handling tasks. The results cross-dataset hyper-parameter sensitivity tests verify generalization ability robustness, respectively. can be considered as an effective solution estimation, thus promoting development field.

Language: Английский

Citations

0