Xijie Jia
Papers
2
Total Citations
29
H-Index
2
About
Xijie Jia is a leading researcher in high-performance hardware acceleration for deep neural networks, with a focus on computer vision applications. His work centers on designing efficient FPGA-based accelerators, particularly on Xilinx’s advanced Versal platform, to overcome the computational and I/O bottlenecks of modern convolutional neural networks (CNNs). Jia’s major contributions include the development of XVDPU, a high-performance CNN accelerator that leverages the Versal AI Engine to achieve superior throughput for 2D vision tasks, and A-U3D, a unified 2D/3D CNN accelerator tailored for disparity estimation in autonomous driving and robotics. These innovations address the growing demands of higher-resolution, more accurate networks by optimizing both computation and data movement. With his most-cited paper, “XVDPU,” garnering 21 citations, and “A-U3D” earning 8, Jia’s work is gaining recognition for enabling practical, real-time deployment of complex vision models. His research bridges the gap between algorithmic advances and hardware efficiency, making him a notable figure in the field of reconfigurable computing and edge AI.
Research Focus
Key Achievements
Top Papers
- 1
- 2