Jinkyu Koo

Papers

Total Citations

H-Index

About

Jinkyu Koo has made significant contributions to the field of computer vision, with a primary focus on efficient deep learning architectures for dense prediction tasks. His most cited work, "Depth Estimation with Simplified Transformer" (2022), addresses a critical challenge in deploying transformer-based models for latency-sensitive applications. By streamlining the transformer architecture, Koo demonstrated that state-of-the-art performance in monocular depth estimation could be achieved without the computational overhead typical of larger models. This work has garnered 11 citations, reflecting its relevance to researchers seeking to balance accuracy and efficiency in real-world vision systems. Koo’s research sits at the intersection of model compression and geometric understanding, offering practical solutions for autonomous navigation and augmented reality. His approach to simplifying complex architectures without sacrificing predictive quality marks a notable achievement in making advanced vision techniques more accessible for deployment. For students and researchers, Koo’s work serves as a valuable case study in designing efficient, high-performing models for resource-constrained environments.