Songyun Deng

Papers

Total Citations

H-Index

About

Songyun Deng is a researcher at the forefront of intelligent robotics and computer vision, with a focused expertise in applying deep learning to automated waste sorting and robotic grasping. His work directly addresses the complex, real-world challenges of kitchen waste management, a domain plagued by high variability in object appearance and a critical lack of annotated datasets. Deng’s major contributions lie in developing cost-effective and data-efficient solutions. His most cited work, "Active Learning-DETR" (2024, 13 citations), pioneers a novel approach to object detection for kitchen waste, integrating active learning to drastically reduce the need for labeled data. He further advances the field with "MCS-ResNet" (2024, 9 citations), a generative grasping network that innovatively fuses RGB and depth information for more robust robotic manipulation. Most recently, his 2025 paper on a multi-objective generative grasping model proposes a complete pipeline for efficient sorting on conveyor belts. Collectively, Deng’s research is notable for its practical impact, bridging the gap between algorithmic innovation and the pressing need for automated, sustainable waste management solutions.