Lihao Jia
Papers
1
Total Citations
25
H-Index
1
About
Lihao Jia is a researcher whose work lies at the intersection of computer vision and machine learning, with a particular focus on egocentric perception and robust learning from imperfect data. His most-cited paper, "Un-supervised and semi-supervised hand segmentation in egocentric images with noisy label learning" (2018), has garnered 25 citations, establishing a foundation for advancing how machines interpret human interaction from first-person perspectives. This work tackles a critical challenge in vision: accurately segmenting hands in wearable camera footage without relying on expensive, cleanly labeled datasets. By introducing novel unsupervised and semi-supervised techniques that learn from noisy labels, Jia's research enables more practical and scalable solutions for applications in augmented reality, human-computer interaction, and assistive robotics. His contributions demonstrate a keen ability to bridge the gap between theoretical machine learning and real-world deployment, offering methods that reduce annotation costs while maintaining high performance. For students and researchers exploring egocentric vision or learning with imperfect supervision, Jia’s work provides both a technical roadmap and an inspiring example of how to address fundamental data challenges in emerging fields.
Research Focus
Key Achievements
Top Papers
- 1