CVRODec 9, 2025

Adaptive Thresholding for Visual Place Recognition using Negative Gaussian Mixture Statistics

arXiv:2512.09071v12025 International Conference on Robotic Computing and Communication (RoboticCC)
Originality Incremental advance
AI Analysis

This addresses a practical challenge in robot navigation by automating threshold selection, though it is incremental as it builds on existing VPR methods.

The paper tackles the problem of automatically selecting thresholds for visual place recognition (VPR) by using negative Gaussian mixture statistics to indicate non-matching images, showing that this approach works well across various image databases and descriptors.

Visual place recognition (VPR) is an important component technology for camera-based mapping and navigation applications. This is a challenging problem because images of the same place may appear quite different for reasons including seasonal changes, weather illumination, structural changes to the environment, as well as transient pedestrian or vehicle traffic. Papers focusing on generating image descriptors for VPR report their results using metrics such as recall@K and ROC curves. However, for a robot implementation, determining which matches are sufficiently good is often reduced to a manually set threshold. And it is difficult to manually select a threshold that will work for a variety of visual scenarios. This paper addresses the problem of automatically selecting a threshold for VPR by looking at the 'negative' Gaussian mixture statistics for a place - image statistics indicating not this place. We show that this approach can be used to select thresholds that work well for a variety of image databases and image descriptors.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes