Zike Yan

Papers

Total Citations

H-Index

About

Zike Yan is a leading researcher at the intersection of computer vision and robotics, with a primary focus on advancing neural implicit representations for 3D scene understanding. Their most significant contribution is the pioneering work on **Continual Neural Mapping** (2021, 34 citations), which introduced a groundbreaking framework for learning implicit scene representations directly from sequential observations. This work addresses a critical limitation in neural radiance fields by enabling models to continuously update and refine their understanding of a scene over time, rather than requiring static, pre-collected datasets. Yan’s research fundamentally bridges the gap between traditional geometry-based SLAM systems and modern deep learning approaches, as demonstrated in their comprehensive survey **Flow-based SLAM** (2019, 12 citations), which systematically examined the evolution from classical geometric computation to learning-based methods. This survey has become an essential reference for researchers transitioning between these paradigms. Yan’s work is particularly notable for tackling the challenge of lifelong learning in neural scene representations, a crucial step toward deploying autonomous systems that can adapt to dynamic, real-world environments. Their contributions continue to influence the development of more robust and adaptive mapping systems for robotics and augmented reality applications.