Jingxiao Zheng
Papers
2
Total Citations
11
H-Index
2
About
Jingxiao Zheng is a leading researcher in energy-efficient domain-specific hardware acceleration for deep learning, with a primary focus on visual object detection and tracking (VODT) for mobile and edge applications. His major contributions lie in designing specialized processors that exploit domain-specific features to dramatically improve energy efficiency compared to general-purpose AI accelerators. His most cited work, "DL-VOPU" (2023, 9 citations), introduces an innovative domain-specific Visual Object Processing Unit that supports multi-scale semantic feature extraction, achieving significant energy savings for tasks like autonomous driving, UAV navigation, and VR/AR. Zheng’s research demonstrates how leveraging the unique characteristics of visual tracking—such as temporal redundancy and fixed-scale processing—can reduce computational overhead without sacrificing accuracy. His work on an energy-efficient visual object tracking processor (2023) further advances this paradigm, showing that domain-specific knowledge is key to overcoming the power constraints of intelligent surveillance and mobile robotics. By bridging the gap between algorithmic efficiency and hardware design, Zheng is enabling real-time, low-power vision intelligence for next-generation autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2