Linhan Wang, a Ph.D. student in computer science, believes that enhancing spatial intelligence represents the next frontier in artificial intelligence.

“This offers a challenge with immense potential to transform how we perceive and engage with our environment,” said Wang, whose research focuses on neural rendering, 3D reconstruction, and the broader realm of 3D computer vision. 

“While there has been remarkable success of large language models in text, code, and math processing,” he said, “these models still struggle to interpret and interact with the 3D world.”

Wang is developing techniques to reconstruct detailed 3D models from sequences of driving videos. 

“By leveraging neural rendering methods, we can generate interactive, novel-view images in real- time technology with significant applications in autonomous driving simulators and virtual reality (VR)/augmented reality (AR) devices,” he said.

Wang holds a bachelor’s degree in atmospheric science from Peking University in China. His current advisor, Chang-Tien Lu, played a pivotal role in his decision to pursue a graduate degree at Virginia Tech because of Lu’s extensive expertise and impressive research record and commitment to mentoring Ph.D. students.

“I appreciate the dynamic, evolving atmosphere at the Sanghani Center, recently enhanced by our move to the Innovation Campus in Alexandria,” he said. “It’s a refreshing new environment and the center’s proactive approach in forging external collaborations makes me feel that I am continuously growing.” 

Among his published papers are “DC-Gaussian: Improving 3D Gaussian Splatting for Reflective Dash Cam Videos,” at the 2024 Neural Information Processing Systems (NeurIPS) conference; and “Self-Correlation and Cross-Correlation Learning for Few-Shot Remote Sensing Image Semantic Segmentation,” at the 2023 International Conference on Advances in Geographic Information Systems 2024 (ACM SIGSPATIAL)

Projected to graduate in 2027, Wang hopes to continue his research in academia or in an industry lab.