Qingan Yan

Papers

Total Citations

222

H-Index

About

Qingan Yan is a leading researcher in 3D computer vision, with a primary focus on point cloud shape completion—a critical task for robotics, autonomous navigation, and augmented reality. His work addresses the fundamental challenge of recovering complete 3D shapes from partial, often noisy, sensor data. Yan’s major contribution lies in developing architectures that preserve fine geometric details during completion, moving beyond earlier methods that produced only coarse, global shapes. His seminal 2020 paper, "Detail Preserved Point Cloud Completion via Separated Feature Aggregation," has garnered 155 citations, establishing a new paradigm by separating global structure from local detail refinement. Building on this, his 2022 work, "Point Cloud Completion Via Skeleton-Detail Transformer" (59 citations), introduced a transformer-based approach that explicitly models the skeleton and surface details, achieving state-of-the-art fidelity. By advancing from global feature encoding to sophisticated, detail-aware frameworks, Yan has significantly improved the realism and utility of completed 3D models, enabling more reliable perception systems. His research continues to shape how machines understand and reconstruct the 3D world.