|
Jingang Shi | Xi'an Jiaotong University I am now a Associate Professor at the School of Software, Xi'an Jiaotong University. Before that, I conducted Postdoc research in CMVS, University of Oulu, where I was advised by Academy Professor Guoying Zhao. During my Postdoc research, I visited Multimedia and Human Understanding Group (MHUG),University of Trento, Italy . Prior to that, I received the B.E. degree and Ph.D. degree from from Xi'an Jiaotong University , China. 入选西安交通大学青年拔尖人才计划,小米青年学者. My research interests include Machine Learning, Human Behaviour Analysis, Emotion AI and Adversarial Learning. I strongly believe we are here, on this planet, time, and realm to experience, learn and grow, which gives me a huge passion about developing myself in a lifetime. My research focuses on advancing multimedia computing through the integration of deep learning, computer vision, and human-centered AI, which mainly contains high-fidelity multimedia restoration, multimodal affective computing, and remote physiological signal monitoring. The work on high-fidelity multimedia restoration addresses real-world degradation models to restore high-fidelity visual content, with applications in digital multimedia preservation, medical imaging, and intelligent surveillance. The research of affective computing develops multimodal emotion recognition frameworks that leverage visual, audio, and physiological signals to enable adaptive human-computer interaction. In addition, the research on remote physiological signal monitoring explores non-contact measurement of vital signs from facial videos under challenging environmental conditions, aiming to enhance healthcare accessibility in daily wellness scenarios. Together, these areas represent a unified effort to design intelligent multimedia systems capable of understanding, restoring, and interpreting complex real-world human-centered data. Attention: We are long-term recruiting self-motivated students to join my research team, focusing on Hybrid Intelligence. Please email me with your CV.
|
|
|
SaTPhys: Sandglass Transformer for Efficient Video-based Remote Physiological Measurement |
|
Reinforcing Transferability and Discriminability for Continual Test-Time Adaptation |
|
Dynamic Recurrent Self-refinement Network for Hyperspectral Remote Sensing Image Super-Resolution |
|
ResoPhys: Unsupervised Plug-and-Play Remote Physiological Measurement via Facial Videos of Arbitrary Resolution |
|
HemNet: Hemoglobin-Assistant Network for Video-based Remote Photoplethysmography Measurement |
|
To remember, to adapt, to preempt: A stable continual test-time adaptation framework for remote physiological measurement in dynamic domain shifts |
|
CodePhys: Robust video-based remote physiological measurement through latent codebook querying |
|
You can wash hands better: Accurate daily handwashing assessment with a smartwatch |
|
MaskFusionNet: A dual-stream fusion model with masked pre-training mechanism for rPPG measurement |
|
Exploiting multi-scale parallel self-attention and local variation via dual-branch transformer-CNN structure for face super-resolution |
|
PhysFormer++: Facial video-based physiological measurement with slowfast temporal difference transformer |
|
Learning Attention from Attention: Efficient Self-Refinement Transformer for Face Super-Resolution |
|
Learning motion-robust remote photoplethysmography through arbitrary resolution videos |
|
PhysFormer: Facial video-based physiological measurement with temporal difference transformer |
|
IDPT: Interconnected Dual Pyramid Transformer for Face Super-Resolution |
|
Atrial fibrillation detection from face videos by fusing subtle variations |
|
Face hallucination via coarse-to-fine recursive kernel regression structure |
|
Hallucinating face image by regularization models in high-resolution feature space |