Zhiyi Shi
Papers
1
Total Citations
25
H-Index
1
About
Zhiyi Shi is a leading researcher in computer vision and remote sensing, with a focus on large-scale aerial scene perception and self-supervised learning. Their most-cited work, "Large-scale aerial scene perception based on self-supervised multi-view stereo via cycled generative adversarial network" (2024, 25 citations), introduces a novel framework that combines multi-view stereo with generative adversarial networks to reconstruct 3D scenes from aerial imagery without requiring labeled data. This contribution addresses critical challenges in autonomous navigation, urban planning, and environmental monitoring by enabling robust perception in data-scarce environments. Shi’s research bridges the gap between synthetic and real-world data through cycled adversarial training, significantly improving depth estimation and scene understanding. With a growing citation impact, their work is recognized for advancing self-supervised methods in geospatial AI, offering scalable solutions for drone-based mapping and disaster response. Shi’s innovative approach to leveraging unlabeled aerial data marks a notable achievement in reducing reliance on expensive manual annotations, positioning them as a rising figure in the intersection of computer vision and remote sensing.
Research Focus
Key Achievements
Top Papers
- 1