Papers
2
Total Citations
178
H-Index
2
About
Xi Peng is a leading researcher in computer vision and deep learning, with a core focus on generative models, multi-view representation learning, and image synthesis. His most impactful contribution is the development of **CR-GAN (Complete Representation GAN)**, a pioneering framework that addresses the fundamental challenge of generating realistic multi-view images from a single input. By identifying that standard GANs often learn "incomplete" representations due to their single-pathway architecture, Peng introduced a dual-pathway learning mechanism that forces the model to capture holistic, view-invariant features. This work, which has accumulated over 178 citations across its versions, has broad applications in robotics, graphics, and autonomous systems where robust 3D understanding from limited 2D data is critical. Peng’s research bridges the gap between generative adversarial networks and practical vision tasks, enabling more reliable object recognition and scene reconstruction under varying viewpoints. His work is widely cited for its theoretical insight into representation completeness and its practical impact on multi-view generation, establishing him as a key innovator in the intersection of generative AI and visual perception.
Research Focus
Key Achievements
Top Papers
- 1CR-GAN: Learning Complete Representations for Multi-view Generation136 citations · 2018
- 2CR-GAN: Learning Complete Representations for Multi-view Generation42 citations · 2018