Weiyue Wang

Papers

Total Citations

199

H-Index

About

Weiyue Wang is a leading researcher in 3D perception for autonomous systems, with a focus on point cloud and range image processing. Their most impactful work, the Sparse Window Transformer (SWFormer), has garnered 126 citations and addresses a fundamental challenge in 3D object detection: the inherent sparsity of point clouds. SWFormer introduces a scalable transformer architecture that efficiently handles sparse 3D data, enabling robust detection critical for robotics and autonomous driving. Wang’s earlier work, "To the Point," achieved 71 citations by pioneering a novel approach that learns 3D representations directly from 2D range images using graph convolution kernels, bridging the gap between efficient 2D processing and accurate 3D understanding. This work demonstrated that spherical coordinates embedded in range images could be leveraged for high-performance detection without costly 3D convolutions. Wang’s contributions are shaping the next generation of perception systems, offering practical solutions for real-world deployment where computational efficiency and accuracy are paramount. Their research continues to influence how autonomous vehicles and robots interpret complex, sparse environments.