Zida Song
Papers
1
Total Citations
330
H-Index
1
About
Zida Song is a leading researcher in computer vision, with a primary focus on the challenging domain of small object detection. His seminal 2020 survey, "A Survey of the Four Pillars for Small Object Detection," has garnered over 330 citations, establishing itself as a foundational reference in the field. In this work, Song systematically defined and analyzed the four critical pillars—multiscale representation, contextual information, super-resolution, and region proposal—that underpin modern small object detection techniques. By synthesizing the limitations of generic object detectors when applied to small, low-resolution targets, he provided a clear roadmap for advancing the state of the art. His contributions are particularly impactful for applications in autonomous driving, surveillance, and remote sensing, where detecting tiny objects is essential. Song’s work not only highlights the persistent challenges posed by limited appearance and geometric cues but also inspires new research directions in deep learning-based detection. For students and researchers entering this field, his survey remains an indispensable starting point for understanding the core problems and solutions in small object detection.
Research Focus
Key Achievements
Top Papers
- 1