Research Projects

Vehicle Occupants Detection. Members: Xiangjian (Sean) He, Wenjing Jia

This research aims to develop new techniques for automatically detecting people in moving vehicles on transit lanes and hence counting the number of vehicle occupants.

We will investigate human/face detection through fusion visible and near-infrared (NIR) images. To improve the performance of the system for detecting people sitting in both front and back seats, images from visible light and NIR light are captured at the same time with two cameras installed at the same spot. The analysis on the fused images is then used to trigger the operation of the unit to take consecutive pictures at the wavelength of NIR light. Then through analysing the multiple frames of NIR images, the occupancy of the vehicles is determined.

Overall, the two major objectives are:
1. Developing new techniques for vehicle classification.
Techniques for recognising the types of vehicles, such as passenger cars, emergency vehicles, motorcycles and bicycles will be developed. A classifier operating on the fused image at the wavelength of visible and NIR light will be trained to determine the type of vehicles.
2. Developing new techniques to detect human faces from NIR images with occlusion.
Unlike face detection from normal images captured at the wavelength of visible light, human faces in NIR images usually do not preserve clear facial features of human. What is more, when the people of interest are sitting in the back seats, successfully detecting them from images will become even more challenging because of the occlusion caused by front-seaters, the front seatbacks and the window frames of vehicles. Through this research, new techniques based on human outline information through analysing multiple consecutive frames of NIR images captured as vehicles passing by the camera will be developed.

The project is funded by the University of Technology, Sydney under the program of Early Career Researcher Grants Scheme (ECRGs).

The key members and the roles of each key member are as follow:

Advisor: Xiangjian (Sean) He, A/Prof.
Project Leader: Wenjing Jia, Dr.
Research Assistant: Damith Mudugamuwa, Dr.

Current Research Projects

	Ad-hoc Location-based Services for Mobile Multimedia. (D. Hoang, J. Chan)
	Cloud Computing and Distributed Systems. (A/Prof. Jinjun Chen, Prof. Doan Hoang, Lingfeng Chen (PhD student), Chang Liu (PhD student), Xuyun Zhang (PhD student), Gaofeng Zhang (PhD student))
	Compressive sensing of time series for human action recognition. (Oscar Perez Concha, Richard Yi Da Xu, Massimo Piccardi)
	Delayed-Multicast for video-on-demand. (D. Hoang, X. He, C. Nguyen)
	Distributed Systems and Service Discovery. (D. Hoang, B. Yousef)
	HMM-MIO: an enhanced hidden Markov model for action recognition. (Ehsan Zare Borzeshi, Oscar Perez Concha, Richard Yi Da Xu, Massimo Piccardi)
	Interactive Visualisation & Visual Analytics of Large Social Networks. (Mao Lin Huang)
	Large Scale Data Visualization. (Mao Lin Huang and Quang Vinh Nguyen)
	Network Congestion Control Modelling. (D. Hoang, M. Li)
	Optical Networking and Control Plane Intelligence. (D. Hoang)
	Peer-to-Peer Overlay Networks. (D. Hoang, H. Le, A. Simmonds)
	Personal Health Monitor. (Peter Leijdekkers and Valrie Gay)
	Programmable/Active Service Platform. (D. Hoang, B. Yousef)
	QoS Mechanisms over Wireless Internet. (D. Hoang, H. Zhou, P. Nhan)
	Scalable End-to-End Policy Routing and Traffic engineering over the Internet. (P. Nanda and A. Simmonds)
	Sign Text Detection from Outdoor Scene. (Xiangjian (Sean) He, Wenjing Jia)
	Spiral Architecture. (Xiangjian (Sean) He, Qiang Wu, Wenjing Jia)
	Vehicle Occupants Detection. (Xiangjian (Sean) He, Wenjing Jia)
	Visual Affective Computing. (Massimo Piccardi, Hatice Gunes)
	Visual Surveillance. (Massimo Piccardi, Hatice Gunes)

Past Research Projects

	Fair Intelligent Quality of Service Control and Path Discovery Mechanisms for Internet Scalable Architectures. (D. Hoang, U. Szewcow, M. Li)
	Fused Data Visualization. (Mao Lin Huang and Wu Quan)