Guanghui He
Papers
1
Total Citations
11
H-Index
1
About
Guanghui He is a leading researcher in efficient deep learning hardware acceleration, with a focus on real-time action recognition and human-computer interaction. His work bridges the gap between complex neural network architectures and practical, low-power deployment on edge devices. He is best known for his pioneering contributions to FPGA-based accelerators for 3D convolutional neural networks, most notably through his highly cited work "AR-C3D: Action Recognition Accelerator for Human-Computer Interaction on FPGA" (2019, 11 citations). In this paper, He proposed a specialized convolutional 3D network optimized for real-time action recognition, significantly reducing computational complexity while maintaining high accuracy. This work has become a foundational reference for researchers developing hardware-efficient solutions for video understanding. Beyond this, He has made substantial contributions to the broader field of embedded AI, demonstrating how custom hardware architectures can enable sophisticated computer vision tasks on resource-constrained platforms. His research continues to influence the design of next-generation intelligent systems for interactive applications, from smart surveillance to assistive robotics.
Research Focus
Key Achievements
Top Papers
- 1