Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Improved Visual Relocalization by Discovering Anchor Points

This repository contains the code for the paper:

Improved Visual Relocalization by Discovering Anchor Points

Soham Saha, Girish Varma, C.V.Jawahar

This paper was accepted as a Conference Paper Spotlight Presentation at BMVC 2018, Newcastle, UK.


Please consider citing our work, if you find it useful in your research:

  title={Improved Visual Relocalization by Discovering Anchor Points},
  author={Saha, Soham and Varma, Girish and Jawahar, CV}


We address the visual relocalization problem of predicting the location and camera orientation or pose (6DOF) of the given input scene. We propose a method based on how humans determine their location using the visible landmarks. We define anchor points uniformly across the route map and propose a deep learning architecture which predicts the most relevant anchor point present in the scene as well as the relative offsets with respect to it. The relevant anchor point need not be the nearest anchor point to the ground truth location, as it might not be visible due to the pose. Hence we propose a multi task loss function, which discovers the relevant anchor point, without needing the ground truth for it.



A model is trained for every scene with different number of anchor points for every scene. The path to the scene and the parameters must be changed in the code.

This code assumes that each scene has a separate folder and is saved in the current path. The paths need to be modified accordingly.


The test performance needs to be computed from the saved files.


Please contact us at :

  • sohamsaha[dot]cs[at]
  • soham[dot]saha[at]research[dot]iiit[dot]ac[dot]in
You can’t perform that action at this time.