Xinwen Zhou

About

Xinwen Zhou is a leading researcher in robotic perception and manipulation, specializing in vision-based grasping and scene understanding. His work addresses fundamental challenges in autonomous robotics, particularly enabling robots to grasp objects in complex, cluttered environments. Zhou's most influential contribution is the development of the **Fully Convolutional Grasp Detection Network with Oriented Anchor Box** (2018), which achieved 226 citations for its real-time, end-to-end approach to predicting multiple grasping poses from RGB images. This work introduced a novel oriented anchor box mechanism and matching strategy, setting a new standard for grasp detection efficiency. He further advanced the field with the **Visual Manipulation Relationship Network** (82 citations), which tackled the critical problem of manipulation in multi-object and object-stacking scenes—a scenario where traditional CNN-based methods fail. Zhou also pioneered the **ROI-based Grasp Detection** approach for overlapping objects, explicitly modeling the affiliation between grasps and their target objects. His research has consistently pushed toward real-time, practical robotic systems, as demonstrated in his work on multi-task CNNs for perception, reasoning, and grasping. With over 460 total citations, Zhou's contributions are essential reading for anyone working in robotic grasping, manipulation, and autonomous systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Fully Convolutional Grasp Detection Network with Oriented Anchor Box

   226 citations · 2018
2. 2
   Visual Manipulation Relationship Network for Autonomous Robotics

   82 citations · 2018
3. 3
   A Real-Time Robotic Grasping Approach With Oriented Anchor Box

   70 citations · 2019
4. 4
   Visual manipulation relationship recognition in object-stacking scenes

   22 citations · 2020
5. 5
   Fully Convolutional Grasp Detection Network with Oriented Anchor Box

   22 citations · 2018
6. 6
   RoI-based Robotic Grasp Detection in Object Overlapping Scenes Using Convolutional Neural Network.

   13 citations · 2018
7. 7
   ROI-based Robotic Grasp Detection for Object Overlapping Scenes

   13 citations · 2019
8. 8
   Visual Manipulation Relationship Detection with Fully Connected CRFs for Autonomous Robotic Grasp

   10 citations · 2018
9. 9
   RPRG: Toward Real-time Robotic Perception, Reasoning and Grasping with One Multi-task Convolutional Neural Network.

   4 citations · 2018
10. 10
    Vision-based Robotic Grasp Success Determination with Convolutional Neural Network

    3 citations · 2017

Key Collaborators

Contact & Links