Scale, Don't Fine-tune: Guiding Multimodal LLMs for Efficient Visual Place Recognition at Test-Time
This work addresses efficiency and generalization issues in visual place recognition for robotics and autonomous systems, representing a novel method for a known bottleneck.
The paper tackles the problem of high computational overhead and limited cross-domain transferability in Visual Place Recognition by proposing a zero-shot framework using Test-Time Scaling, which achieves up to 210x computational efficiency gains and superior generalization across diverse environments.
Visual Place Recognition (VPR) has evolved from handcrafted descriptors to deep learning approaches, yet significant challenges remain. Current approaches, including Vision Foundation Models (VFMs) and Multimodal Large Language Models (MLLMs), enhance semantic understanding but suffer from high computational overhead and limited cross-domain transferability when fine-tuned. To address these limitations, we propose a novel zero-shot framework employing Test-Time Scaling (TTS) that leverages MLLMs' vision-language alignment capabilities through Guidance-based methods for direct similarity scoring. Our approach eliminates two-stage processing by employing structured prompts that generate length-controllable JSON outputs. The TTS framework with Uncertainty-Aware Self-Consistency (UASC) enables real-time adaptation without additional training costs, achieving superior generalization across diverse environments. Experimental results demonstrate significant improvements in cross-domain VPR performance with up to 210$\times$ computational efficiency gains.