Papers
2
Total Citations
8
H-Index
2
About
Marek Cygan is a roboticist whose research lies at the intersection of deep learning and robotic manipulation, with a primary focus on improving autonomous grasping systems. His most recognized contribution, the improved Grasp Quality Convolutional Neural Network (GQ-CNN), builds upon the DexNet 2.0 dataset to significantly enhance grasp success rates for unknown objects—a critical challenge in industrial and service robotics. This work, published in 2018, has accumulated 6 citations and represents a key step toward more reliable robot perception. More recently, Cygan has advanced the field with his "Grasping Student" framework (2023), which introduces a semi-supervised learning approach to robotic manipulation. By leveraging product images alongside a small set of real-world robot experiences, this method directly addresses the data bottleneck that often limits the deployment of learning-based systems in practical settings. Though early in its citation lifecycle, this work signals a promising direction for scalable, data-efficient robot learning. Cygan’s research is particularly valuable for students and engineers seeking to bridge the gap between simulation-trained models and real-world robotic performance.
Research Focus
Key Achievements
Top Papers
- 1Improved GQ-CNN: Deep Learning Model for Planning Robust Grasps6 citations · 2018
- 2Grasping Student: semi-supervised learning for robotic manipulation2 citations · 2023