Kazuki Kozuka
Papers
1
Total Citations
2
H-Index
1
About
Kazuki Kozuka’s research lies at the intersection of computer vision, human activity recognition, and assistive robotics, with a particular focus on understanding complex, multi-person interactions in everyday environments. His most cited work, “Few-Shot Classification of Interactive Activities of Daily Living (InteractADL)” (2024), addresses a critical gap in the field: while most benchmarks focus on single-person tasks, Kozuka’s research pioneers methods for recognizing interactive ADLs—such as cooking together or assisting someone—that are essential for smart homes, healthcare, and assistive robots. By leveraging few-shot learning techniques, his work enables models to generalize from limited examples, a practical necessity for real-world deployment. Though his citation count is still growing, the InteractADL benchmark has already garnered attention for its novel focus on multi-person dynamics, setting a new direction for activity recognition research. Kozuka’s contributions are particularly impactful for developing robots that can understand and assist in collaborative human tasks, making his work a cornerstone for future advances in human-robot interaction and ambient assisted living.
Research Focus
Key Achievements
Top Papers
- 1