This repo contains VPR models that have been fine-tuned for indoor usage. Currently available models come from the following repositories:
- EigenPlaces: Training Viewpoint Robust Models for Visual Place Recognition;
- Optimal Transport Aggregation for Visual Place Recognition (Salad).
The fine-tuned procedure is inspired by Deep Visual Geo-localization Benchmark
The indoor datasets used for fine-tuning are: