Brendan McCane

Papers

Total Citations

173

H-Index

About

Brendan McCane is a leading researcher in computer vision and robotics, with key contributions spanning scene understanding, 3D data compression, and reinforcement learning. His most cited work, "SIFT and SURF Performance Evaluation against Various Image Deformations on Benchmark Dataset" (2011, 166 citations), provides a foundational analysis of feature invariance for scene classification—a critical problem for autonomous systems navigating indoor and outdoor environments. This study remains a go-to reference for selecting robust image features under transformations like rotation, illumination, and viewpoint changes. McCane also advances 3D scene sensing through "Variational Autoencoder for 3D Voxel Compression" (2020), addressing the storage challenges of voxel grids for lightweight robotics applications. More recently, his work "Learning to explore by reinforcement over high-level options" (2023) explores hierarchical reinforcement learning, enabling agents to efficiently navigate complex tasks. McCane’s research bridges theoretical rigor and practical deployment, with his citation impact underscoring its relevance to both vision and robotics communities. His contributions empower students and researchers tackling real-world perception and decision-making challenges.