Papers
8
Total Citations
169
H-Index
5
About
Dehao Huang is a leading researcher in robotic manipulation, with a primary focus on **task-oriented grasping (TOG)**—the challenge of enabling robots to grasp objects in a way that is compatible with a subsequent manipulation task. His work sits at the intersection of **computer vision, natural language processing, and robotics**, pioneering the integration of large language models and foundation models into grasping pipelines. Huang’s most impactful contribution is **GraspGPT** (2023, 88 citations), which leverages semantic knowledge from large language models to predict task-appropriate grasps, setting a new standard for TOG. He further advanced the field with **FoundationGrasp** (2025, 22 citations), demonstrating how foundation models can generalize grasping strategies across diverse objects and tasks without task-specific training. His work on **visual-language inputs for grasp prediction** (38 citations) bridges the gap between human commands and robotic action by jointly solving object grounding and task grounding. Beyond grasping, Huang has contributed to **cloud-based VSLAM** and **efficient object rearrangement via multi-view fusion**, showcasing a broader interest in autonomous navigation and scene understanding. With over 160 citations and a string of recent publications (2023–2025), Dehao Huang is at the forefront of making assistive robots that can understand language, reason about tasks, and manipulate tools intelligently in human environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2Task-Oriented Grasp Prediction with Visual-Language Inputs38 citations · 2023
- 3FoundationGrasp: Generalizable Task-Oriented Grasping With Foundation Models22 citations · 2025
- 4
- 5Efficient Object Rearrangement via Multi-view Fusion6 citations · 2024
- 6
- 7Task-Oriented Grasp Prediction with Visual-Language Inputs2 citations · 2023
- 8