Papers
5
Total Citations
91
H-Index
4
About
Wenjie Geng is a leading researcher in robotic manipulation, with a primary focus on vision-based grasp detection for complex, real-world environments. His work addresses a critical challenge in robotics: enabling robots to reliably grasp objects despite background clutter, occlusion, and other visual disturbances. Geng’s major contributions include the development of novel deep learning architectures, such as a two-stream convolutional neural network (CNN) that simultaneously performs object detection and segmentation for robotic grasping, and a vision-based method that combines an SSD detector with an image inpainting and recognition network to handle occlusion. His research has garnered significant attention, with his most cited paper, "A Two-Stream CNN With Simultaneous Detection and Segmentation for Robotic Grasping," accumulating 38 citations since 2020. He has also advanced pixel-wise grasp detection through innovative techniques like twin deconvolution and multi-dimensional attention to mitigate checkerboard artifacts, and hierarchical multi-scale feature fusion for improved accuracy. Geng’s work is highly influential in the field of robotic manipulation, providing practical solutions for deploying robots in disturbed and occluded scenes, and his methods are foundational for future research in autonomous grasping systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5