Build a Deep Learning Model for Camera Relocalization

Job ID: 34151583

Budget: $250 – $750 USD

Hello,

I am looking for someone with expertise in Deep Learning, Computer Vision, and Robotics (Lidar, Structure from motion).

The project is about camera relocalization (https://paperswithcode.com/task/camera-relocalization) using deep learning. Given a set of images I want to be able to get the camera poses.

Here is the dataset: https://www.naverlabs.com/en/storyDetail/211

Here are the Deep Learning models that need to be implemented on the above dataset :

DL 1 : NetVLAD + R2D2 (https://paperswithcode.com/paper/large-scale-localization-datasets-in-crowded)
DL 2 : Patch-NetVLAD + R2D2 (https://paperswithcode.com/paper/patch-netvlad-multi-scale-fusion-of-locally)
DL 3 : https://github.com/XT-1997/FDANET

Please do not contact me before checking these requirements in detail.