Research Areas

Computer Vision and Pattern Recognition (CVPR)

The CVPR research group focuses on developing the state-of-the-art models, theories and core technologies for facial recognition, biometric system security medical informatics, medical image processing and video surveillance. The research group not only aims to develop enabling technologies of all these applications, but also to address the public concerns of security and privacy issues.

Faculty Involved:

• Dr. CHEN, Jie
• Prof. CHEUNG, Yiu Ming
• Dr. WAN, Renjie
• Prof. YUEN, Pong Chi
• Dr. ZHOU, Kaiyang

Funded Research and Consultancy Projects in the Past Few Years:

Interpretable Machine Learning Aided Understanding of Complex Brain Structural and Functional Interaction Underlying Spectra of Individual Differences in Cognitive Behaviour
	Staff	Prof. CHEUNG, Yiu Ming
	Abstract	To develop a core set of machine and deep learning models and tools for big brain data To gain new understanding of how the brain employ specific systems for selected social cognition tasks To further develop and apply methods to assess autistic trait

On Developing a Lip-password Based Face Recognition System
	Staff	Prof. CHEUNG, Yiu Ming
	Abstract	To integrate lip-password into the face recognition system as a single learning paradigm; To select and design the underlying models and algorithms of building up the lip-password-based face recognition system.

A Hybrid Residual Network and Long Short-Term Memory Method for Peptic Ulcer Bleeding Mortality Prediction
	Staff	Prof. YUEN, Pong Chi
	Abstract	To design a hybrid residual network and long short-term memory method for accurate peptic ulcer bleeding mortality prediction.

Face anti-spoofing to combat mask attacks: A remote photoplethysmography approach
	Staff	Prof. YUEN, Pong Chi
	Abstract	To effectively detect the 3D mask attack with different mask materials and qualities, a new liveness cue is proposed by analyzing heartbeat signal through remote photoplethysmography (rPPG). To precisely identify the heartbeat vestige from the observed noisy rPPG signals under practical lighting conditions, A rPPG correspondence feature for 3D mask face anti-spoofing is proposed.

Multi-adversarial Discriminative Deep Domain Generalization for Face Presentation Attack Detection
	Staff	Prof. YUEN, Pong Chi
	Abstract	This work focuses on improving the generalization ability of face anti-spoofing methods from the perspective of the domain generalization. Learning a generalized feature space via a novel multi-adversarial discriminative deep domain generalization framework.

Towards Practical Object Trackers: From Feature Combination to Modality Fusion
	Staff	Prof. YUEN, Pong Chi
	Abstract	Developing robust visual tracking and object re-detection framework based on information fusion techniques to handle large appearance variations and tracking loss Developing feature combination models which can adaptively combine appropriate visual features Developing modality fusion models which can exploit the complementarity of other non-visual modalities (data sources) for appearance modeling

Unsupervised Embedding Learning via Invariant and Spreading Instance Feature
	Staff	Prof. YUEN, Pong Chi
	Abstract	Learning a discriminative feature extraction Deep Neural Network (DNN) with large-scale unlabeled images, such that the visually similar samples are close to each other in the learned embedding space