An ensemble learner consisting of a SVM classifier and active learning for video analysis and representation is proposed. This ensemble learner would learn the structure of the captured video footage that is input as the training data set. It continues to learn video structure as new examples are input. In this manner, the learner is expected to achieve the following:

1. Quickly establish the query concept, and
2. Accurately identify and learn video features from a few labelled instances

This project also proposes to investigate the appropriateness of applying different kernel methods to improve overall performance of the ensemble learner. Results of video data learning would enable collection of video metadata, and would also be another step towards automated video data indexing. This is yet closer to facilitating the Semantic Web and its services.

Research by Dr Adelina Tang and Mr Phua Yeong Tsann

The highlight of this research work is radio wave propagation in urban drains environment (cities such as Sunway, or other new township). The drainage systems in the US and Europe differ notably from those in the South East Asian countries like Malaysia in that the former are primarily below ground and covered while the latter are partially open on the top. We are interested to investigate how different these two environments differ from each other and whether or not the existence of these partially open drains support radio propagation in small radio cells, which are highly relevant in Malaysian cities.

Research by Dr Grace Lim Soo Yong

An old native Ameraican proverb said “Tell me and I’ll forget, show me and I may remember, involve me and I’ll understand”. It will be more effective to broadcast messages to people using interactive announcement display walls that attract and involve the interests of passers-by. Instead of static displays or basic interactive kiosks, the display wall detects the user’s position and projects dynamic imagery annoucements based on their motion to create an engaging and informative environment. The proposed system enables the user to interact with the contents on a large display wall using vision-based techniques. The system uses camera to recognise the position of the body or hands when the user is standing in front of the display wall. Based on body tracking and recognition information, the contents are changed accordingly to interact with the user.

Research by Dr Lee Yun Li

This research aims to develop a positioning method for in-network processing in collaborative and distributed wireless network. The design requirement is to improve the accuracy and precision of indoor location estimation via radio ranging. To evaluate and demonstrate the achivement, we expect to apply the research output for indoor navigation applications using an Android smartphone.

Research by Dr Pu Chuan Chin

The Malaysian government has outlined the expansion and development of communication infrastructure as one of the twelve National Key Economic Areas of the Economic Transformation Program. In line with this effort, researchers from Sunway University has started a major industrial based research project on Rural Wireless Broadband Network as the enabling technology to provide wireless broadband access to the rural population. The deployment of wireless broadband in Malaysia rural areas has complex effects across various domains including social, technological, economic, environmental, and political, and hence our experts ranging from technology forecasting to network simulation to wireless platform development have been actively involved to conduct research on the challenges brought about by this project. The outcome is: a next generation wireless broadband access system tailor-made to rural Malaysia.

Research by Dr Yau Kok Lim