Chia-Yu Chang

Papers

Total Citations

H-Index

About

Chia-Yu Chang is a leading researcher in energy-efficient deep learning hardware, with a primary focus on convolutional neural network (CNN) processors for real-time computer vision. Their most impactful work centers on advancing object detection in high-resolution video streams, particularly for safety-critical applications like autonomous driving and advanced driver-assistance systems (ADAS). Chang’s major contribution is the development of a 16nm CNN processor achieving 5.7 TOPS, which uniquely supports a bi-directional feature pyramid network (FPN) for enhanced small-object detection. This innovation is vital for enabling autonomous vehicles to perceive distant obstacles, ensuring safe following distances and potentially saving lives. The processor’s design addresses the extreme computational demands of high-resolution video, balancing performance with power efficiency. While their most cited paper is recent (2025), its technical novelty in integrating bi-directional FPN into a dedicated chip marks a significant step toward practical, real-time small-object detection in edge devices. Chang’s work sits at the intersection of computer architecture and computer vision, promising to make intelligent systems more responsive and reliable in dynamic environments.