Peiji Li
Papers
2
Total Citations
14
H-Index
2
About
Peiji Li is a researcher focused on advancing scene text recognition, particularly for challenging, real-world applications in the logistics industry. Their major contributions center on developing robust models capable of accurately interpreting low-quality, curved, and distorted text—such as that found on express sheets and shipping labels. Li’s work directly addresses the gap between laboratory-perfect text recognition and the messy reality of industrial deployment. Notably, their 2022 paper on context modeling for low-resolution logistics images has garnered 12 citations, demonstrating its foundational impact. Building on this, Li introduced MTSTR (Multi-Task Learning for Scene Text Recognition), a novel framework employing a dual attention mechanism to further enhance performance on degraded inputs. This multi-task approach, published in 2023, represents a significant step toward integrating text recognition into complex systems like robot vision and automated logistics sorting. By tackling the specific, high-stakes challenges of the logistics industry, Peiji Li is helping to bridge the gap between computer vision research and practical, large-scale industrial intelligence.
Research Focus
Key Achievements
Top Papers
- 1
- 2