Zakaria Laskar* · Iaroslav Melekhov* · Assia Benbihi · Shuzhe Wang · Juho Kannala
We propose a hybrid scene compression method, D-PQED, which performs descriptor quantization-dequantization in an end-to-end differentiable manner. This approach is well-suited for structure-based localization methods, enabling accurate camera pose prediction under a very limited memory budget.
If you find our code or paper useful, please cite
@inproceedings{Laskar2024dpqed,
author = {Laskar, Zakaria and Melekhov, Iaroslav and Benbihi, Assia and Wang, Shuzhe and Kannala, Juho},
title = {Differentiable Product Quantization for Memory Efficient Camera Relocalization},
journal = {European Conference on Computer Vision (ECCV)},
year = {2024},
}The Readme template was mainly inspired by the MonoSDF repository.