Jiquan Ngiam

Papers

Total Citations

189

H-Index

About

Jiquan Ngiam is a researcher specializing in 3D object detection and perception systems for autonomous driving and robotics. His work sits at the intersection of deep learning and spatial understanding, with a particular focus on developing efficient, accurate methods for interpreting three-dimensional environments from sensor data. Ngiam's most recognized contribution is the **3D Multi-frame Attention Network (3D-MAN)**, a 2021 architecture that addresses a critical limitation in existing detection systems — the over-reliance on single-frame data. By leveraging attention mechanisms across multiple temporal frames, 3D-MAN achieves richer scene understanding, accumulating over 114 citations and establishing itself as an influential reference in the autonomous perception community. His complementary work, "To the Point," further demonstrates his commitment to practical efficiency. By learning 3D representations directly from perspective range images using graph convolution kernels, this approach offers a computationally elegant alternative to volumetric methods, earning 71 citations and highlighting Ngiam's ability to bridge theoretical innovation with real-world deployment constraints. Together, these contributions reflect a research vision centered on making 3D perception more robust, temporally aware, and computationally feasible — qualities essential for the next generation of autonomous systems.