Papers
1
Total Citations
28
H-Index
1
About
G. Csurka is a leading researcher in computer vision, with key contributions to 3D reconstruction, image matching, and visual recognition. Her most-cited work, "Finding the Collineation between Two Projective Reconstructions" (1999, 28 citations), introduced a foundational method for aligning projective reconstructions—a critical step in multi-view geometry that underpins modern structure-from-motion pipelines. This work, co-authored with luminaries like Andrew Zisserman, demonstrated how to compute collineations between uncalibrated views, enabling robust 3D scene recovery from image sequences. Beyond this, Csurka’s research spans visual categorization, where she pioneered bag-of-words models for image classification, and domain adaptation, advancing techniques for transferring knowledge across visual domains. Her impact is reflected in the enduring relevance of her geometric methods, which continue to influence autonomous navigation, augmented reality, and photogrammetry. Csurka’s ability to bridge theoretical geometry with practical vision systems has made her a respected figure in the field, inspiring students and researchers to explore the intersection of algebraic methods and real-world visual understanding.
Research Focus
Key Achievements
Top Papers
- 1Finding the Collineation between Two Projective Reconstructions28 citations · 1999