Yukang Gan

About

Yukang Gan is a leading researcher in computer vision and perceptual robotics, whose work has significantly advanced RGB-D scene understanding. His primary research areas include deep learning for semantic scene labeling, multimodal data fusion, and context-aware modeling. Gan’s most impactful contribution is the development of LSTM-CF, a pioneering framework that unifies context modeling and feature fusion using Long Short-Term Memory networks for RGB-D scene labeling. This work, which has garnered 188 citations, introduced a novel approach to generating pixelwise, fine-grained label maps by simultaneously processing photometric (RGB) and depth channels. The method addresses a critical challenge in perceptual robotics: enabling machines to understand complex, cluttered environments by leveraging both visual and spatial information. Gan’s research has been instrumental in improving the accuracy and robustness of scene parsing, directly benefiting applications in autonomous navigation and robotic manipulation. His 2016 paper on the Long Short-Term Memorized Fusion Model further refined these techniques, demonstrating his sustained focus on integrating temporal and spatial contexts. Through these contributions, Gan has established himself as a key figure in advancing intelligent systems that can perceive and interact with the physical world.

Research Focus

Key Achievements

Top Papers

1. 1
   LSTM-CF: Unifying Context Modeling and Fusion with LSTMs for RGB-D Scene Labeling

   188 citations · 2016
2. 2
   RGB-D Scene Labeling with Long Short-Term Memorized Fusion Model.

   12 citations · 2016

Key Collaborators

Contact & Links