Guoying Zhao - Selected Publications#
Publications: 3 books, 6 book chapters, and over 300 papers. Citations: GS h-index 72, i10-index 199, citations 23253+ (May 2023). Most cited: PAMI 2007: 3168+ citations. For recent five years, the average citation number is 3100+/year.
1) X. Li, X. Hong, A. Moilanen, X. Huang, T. Pfister, G. Zhao, and M. Pietikäinen. Towards Reading Hidden Emotions: A Comparative Study of Spontaneous Micro-expression Spotting and Recognition Methods. IEEE Transactions on Affective Computing, 9(4): 563-577, 2018. (296+ citations/Google Scholar)
This paper proposed the first automatic complete ME analysis system that can not only detect but also recognize MEs from spontaneous video data. The method outperforms humans in ME recognition by a large margin, and achieves comparable performance to humans in the very challenging task of both detecting and recognizing spontaneous MEs.
This work was reported by MIT Techonology Review, stating “emerging technology…it’s easy to imagine this being used in …assessments and even in Google Glass-type devices”, later widely reported in the public press, e.g., Daily Mail and Technology Discovery TV.
2) W. Peng, X. Hong, H. Chen, G. Zhao. Learning Graph Convolutional Network for Skeleton-based Human Action Recognition by Neural Searching. Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI 2020).
(202+ citations/Google Scholar for two years; The shared Github code implementation has obtained 140+ stars.)
Proposed to use Neural Architecture Search (NAS) for the the first automatically designed GCN for skeleton-based action recognition.
3) Z. Yu, W.Peng, X. Li, X. Hong, G. Zhao. Remote Heart Rate Measurement from Highly Compressed Facial Videos: an End-to-end Deep Learning Solution with Video Enhancement. ICCV 2019.
(151+ citations/Google Scholar for two years)
Propose a two-stage, end-to-end method using hidden rPPG information enhancement and attention networks, which is the first attempt to counter video compression loss and recover rPPG signals from highly compressed videos. This work was awarded IEEE Finland section best student paper.
4) Z. Yu, C. Zhao, Z. Wang, Y. Qin, Z. Su, X. Li, F. Zhou, G. Zhao. Searching Central Difference Convolutional Networks for Face Anti-Spoofing. CVPR 2020.
(200+ citations/Google Scholar) .
Proposed a novel frame level face anti-spoofing method based on Central Difference Convolution (CDC), which is able to capture intrinsic detailed patterns via aggregating both intensity and gradient information. With the method extended from this paper, we won the first place on Chalearn multi-modal face anti-spoofing attack detection challenge @ CVPR 2020.
5) Y. Li, X. Huang, G. Zhao. Micro-expression Action Unit Detection with Spatial and Channel Attention. Neurocomputing. 2021.
It is the first work on micro-expression action unit detection, providing the opportunity to analyze the mapping of facial action unit to hidden emotions. It is now being followed widely and facial micro-expression action unit detection has been recognized more suitable for largely reducing the the consistency of the labels. and better analyzing emotions.
6) Z. Yu, J. Wan, Y. Qin, X. Li, S.Z. Li & G. Zhao. NAS-FAS: Static-Dynamic Central Difference Network Search for Face Anti-Spoofing. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). 2021. (Impact factor: 16.39)
This is the first work on face anti-spoofing with neural architecture search and published in one of the best journal in CV and PR fields.
7) J. Shi, I. Alikhani, X. Li, Z. Yu, T. Seppänen & G. Zhao. Atrial Fibrillation Detection from Face Video by Fusing Subtle Variations. IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 30 (8): 2781-2795, 2020.
It is the first work of Atrial Fibrillation detection with remote face video analysis, opening new potentials for the application of remote heart rate measuring to medical field.
8) X. Liu, H. Shi, H. Chen, Z. Yu, X. Li, G. Zhao. iMiGUE: An Identity-free Video Dataset for Micro-Gesture Understanding and Emotion Analysis. CVPR, 2021.
It is the first micro-gesture work with the first in the wild dataset. Within one year after publication, the dataset has bee requested and shared to 20+ groups in the world.
9) G. Zhao and M. Pietikäinen. Dynamic Texture Recognition Using Local Binary Patterns with an Application to Facial Expressions. IEEE Transactions on Pattern Analysis and Machine Intelligence journal (TPAMI), 2007, 29(6): 915-928. (3070+ citations/GS)
This publication in a top journal in computer vision proposed two novel spatiotemporal descriptors, VLBP and LBP-TOP, which for the first time considered appearance and motion in videos in the form of local texture patterns and formulated the co-occurrence of local patterns in both spatial and temporal domains, achieving superior performance on both dynamic texture and facial expression recognition experiments. This work led to many other teams around the world investigating the approach, proposing many LBP-TOP variants and using the method for baseline evaluation in many public challenges and competitions, including Emotion Recognition In The Wild Challenges (2013, 2014, 2015, 2016, 2017), Facial Micro-Expression Grand Challenge (2018, 2019), and the Multimodal Emotion Recognition Challenge (2016).
10) H. Chen, H. Shi, X. Liu, X. Li and G. Zhao. SMG: A Micro-Gesture Dataset Towards Spontaneous Body Gestures for Emotional Stress State Analysis. International Journal on Computer Vision (IJCV).
This paper tackles the problem of micro-gesture (MG) analysis, which is an understudied problem in the computer vision community. A new dataset of micro gestures was created and annotated, which is first of its kind with low level to high level annotations. Novel frameworks are presented together with various state-of-the-art methods as benchmarks for automatic classification of MG, online recognition, and emotional stress state recognition from MGs.
It was accepted in the first round of review with recommendation as accept with minor revision, to this top level journal in computer vision field, with the comments such as "The paper can be considered a pioneering work and an important step towards solving a newly identified problem in computer vision - MG detection and analysis to infer hidden emotional state of people. As such it would be a valuable contribution to the IJCV".
Zhao’s work has been leading to new research directions such as facial micro-expression analysis, micro-gesture recognition, and carrying out pioneering research in remote physiological signal measure from videos, face anti-spoofing, human behavior understanding with multi-disciplinary application study in security, education and health.
