$ show gmberton/VPR-methods-evaluation
This repository is a research evaluation harness for Visual Place Recognition (VPR) models. It wraps a wide range of state-of-the-art VPR methods behind a common CLI, allowing users to run any supported model on standard VPR datasets (e.g., St Lucia) or a bundled toy dataset. The main entry point is test_dataset.py, which loads database/query image folders, extracts descriptors using the selected model, computes recall@N metrics, and saves logs and optional visualizations. The repo is designed to work with the companion VPR-datasets-downloader to standardize dataset format. Although it is a CLI tool, its primary purpose is research evaluation, so it is classified as data_research.
The LLM identified 3 files worth highlighting for reuse:
PyTorch Dataset for VPR evaluation that reads database/query image paths from text files or directories, parses UTM coordinates embedded in filenames to generate ground-truth positives, and supports flexible image loading. This is a self-contained data-loading utility that can be extracted and reused for any VPR dataset following the same path/UTM convention.
Utility function that writes a query image path, its predicted database matches, and ground-truth positives to a text file. This enables error analysis and visualization of VPR predictions. It is decoupled from the model and can be reused in any retrieval evaluation pipeline.
Central registry mapping VPR model names to constructors and metadata (descriptor dimension, image size, etc.). It decouples the evaluation script from model-specific implementations, making it easy to add new methods by registering them. This is a clean example of a plugin/factory pattern for research codebases.