Oliver Grainge

University of Essex, University of Southampton

Papers

Total Citations

H-Index

About

Oliver Grainge is a researcher advancing the frontier of efficient visual place recognition (VPR) for robotics and autonomous systems. His work focuses on making high-accuracy VPR models deployable on resource-constrained devices through model compression and quantization. Grainge’s major contributions include pioneering the use of low-bit quantized neural networks for VPR, demonstrating that aggressive quantization can maintain robust localization performance under challenging conditions like illumination changes and occlusion. His most cited paper, “Design Space Exploration of Low-Bit Quantized Neural Networks for Visual Place Recognition” (2024, 7 citations), systematically evaluates trade-offs between bit-width and accuracy, providing a practical guide for embedded deployment. He further advanced the field with “TeTRA-VPR” (2025), introducing a ternary transformer approach that achieves state-of-the-art compactness, and “Structured Pruning for Efficient Visual Place Recognition” (2024), which reduces model size without sacrificing recognition capability. Grainge’s work directly addresses the critical need for lightweight, real-time VPR in robotics, enabling global re-localization on devices with limited memory and compute. His research is foundational for students and engineers seeking to bridge the gap between high-performance vision transformers and practical, on-device navigation systems.