Research Associate - Object Detection & Segmentation

About the role

The Hyundai-NTU-A*STAR Corp Lab at NTU's Main Campus is running a robotic pick and place project and is looking for a research associate to handle its computer vision component. Only shortlisted candidates will be notified.

Key responsibilities

• Develop 3D object detection and segmentation models for the robotic pick and place project.
• Develop data augmentation and data synthesis methods to cope with limited training data.
• Develop algorithms to annotate 2D images using 3D images.
• Check data quality and review existing 3D object detection and segmentation methods, identifying their weaknesses.
• Carry out experiments, analyse results and prepare reports and scientific papers.
• Work with other researchers and engineers on the robotic pick and place system.

Required profile

• Master's degree in Computer Engineering, Computer Science, Electronics Engineering or equivalent.
• More than 2 years of research experience preferred.
• Strong research capabilities: innovative thinking, hypothesis formulation, experimental design and project reporting.
• Collaborative and communication skills, with a proactive, analytical mindset and the ability to work independently or in a team.

Required skills

• In-depth knowledge of deep learning, including 2D and 3D object detection and segmentation, SAM, foundation models, diffusion models, transfer learning and few-shot object detection and segmentation, and 3D computer vision.
• Publications at top AI conferences such as CVPR, ICCV, ECCV, NeurIPS, ICML and ICRL preferred.
• Preferred: software implementation expertise and knowledge of robotics.