UNIMY Secures FRGS Grant to Advance Next-Generation AI for Smart Cities

UNIMY is proud to announce that our researchers have secured a Fundamental Research Grant Scheme (FRGS) award worth RM97,400 from the Ministry of Higher Education (MOHE) for the project “Graph-based Dynamic Edge Strategy for Human–Object Interaction Recognition.”

Led by Prof. Dr. Habibollah Haron with co-researchers Noor Fadzillah Azman and Nur Faiqah Ab Hamid@Fauzi, the project positions UNIMY at the forefront of AI and computer vision innovation.

At its core, the research tackles a key challenge in intelligent systems, which is how computers can reliably understand what people are doing with objects around them - whether lifting a box, using a tool, or interacting with equipment in public spaces. Existing Human–Object Interaction (HOI) models struggle with accuracy and require heavy computing power.

This project responds by designing three cutting-edge components, namely the Interaction Recognition Graph Neural Network (IR-GNN) to filter out non-relevant human– object pairs, the Parallel Multi Head Graph Attention Network (PMGAT) to focus on fine details such as body posture and facial cues and a Graph Sampling based Dynamic Edge Strategy (GraphSADES) to make the whole system faster and more efficient.

Using benchmark datasets, the team will build and test models that can deliver higher precision with lower computational cost. The outcomes will support safer roads through better driver-monitoring systems, smarter surveillance for public security and more responsive human– machine collaboration in areas such as healthcare, robotics and industrial automation.

Beyond publications and potential IP, the project will nurture postgraduate talent and provide rich research problems for UNIMY students who are keen to work in AI, data science and smart city solutions. This FRGS success underscores UNIMY’s commitment to advancing Malaysia’s digital economy and contributing AI technologies that are not only powerful, but purposeful and socially relevant.