Bingce Du
Papers
1
Total Citations
56
H-Index
1
About
Bingce Du is a researcher specializing in efficient deep learning architectures, with a particular focus on real-time semantic segmentation for computer vision. Their most significant contribution is the development of LAANet (Lightweight Attention-Guided Asymmetric Network), introduced in a 2022 paper that has garnered 56 citations. This work addresses a critical challenge in deploying neural networks on resource-constrained devices: achieving high segmentation accuracy while maintaining real-time performance. By designing an asymmetric encoder-decoder structure guided by attention mechanisms, Du's LAANet balances computational efficiency with spatial detail preservation, making it highly relevant for autonomous driving, robotics, and mobile applications. The paper's citation count reflects its impact as a practical solution for edge deployment, bridging the gap between heavy, high-accuracy models and lightweight, fast alternatives. Du's research contributes to the broader goal of making advanced vision models accessible for real-world, latency-sensitive tasks, positioning them as a notable voice in the field of efficient deep learning.
Research Focus
Key Achievements
Top Papers
- 1