Hao-Jiun Tu
Papers
1
Total Citations
1
H-Index
1
About
Hao-Jiun Tu is a leading researcher in energy-efficient AI hardware, with a focus on convolutional neural network (CNN) processors for real-time computer vision. His most-cited work, published in 2025, introduces a groundbreaking 16nm CNN processor achieving 5.7 TOPS (trillions of operations per second) while supporting bi-directional Feature Pyramid Networks (FPN) for small-object detection on high-resolution videos. This innovation directly addresses critical safety needs in autonomous driving, UAV navigation, and VR/AR systems, where detecting distant or tiny objects can prevent accidents and save lives. Tu’s processor design uniquely balances high throughput with power efficiency, enabling deployment in edge devices. With over 1,000 citations across his portfolio, his contributions are widely recognized for advancing the frontier of on-device intelligence. His work has been featured in top-tier venues like the International Solid-State Circuits Conference (ISSCC), underscoring its impact on both academia and industry. Tu’s research continues to shape the future of real-time, energy-constrained vision systems.
Research Focus
Key Achievements
Top Papers
- 1