search

ethz-asl / maplab

A Modular and Multi-Modal Mapping Framework
https://maplab.asl.ethz.ch
Apache License 2.0
2.55k stars 722 forks source link

Can I use landmark sparsification using objective function of quadratic term #179

Open giho-sung opened 4 years ago

giho-sung commented 4 years ago

Hello,

I appreciate your works! I want to say it before a question.

I'm trying to use map sparsification function in [Keep it Brief: Scalable Creation of Compressed Localization Maps]. The objective functions it uses are consist of linear terms and a quadratic term. I want to use those objective functions, but a linear objective function seems available in maplab. https://github.com/ethz-asl/maplab/blob/88d8587958c6902e7304a5ef7d88e867b30924b3/algorithms/map-sparsification/src/optimization/lp-solve-sparsification.cc#L212-L236 I found some code on quadratic term. https://github.com/ethz-asl/maplab/blob/88d8587958c6902e7304a5ef7d88e867b30924b3/algorithms/map-sparsification/src/optimization/quadratic-term.cc But, it looks a stand-alone code now. How can I use an objective function this paper use?

Thanks,

Giho.

dymczykm commented 4 years ago

@sa1522 thanks for your question.

The problem is that the public version of maplab uses "lpsolve" which doesn't support quadratic terms so we couldn't port our algorithms to this optimization framework. Originally, we used Gurobi optimizer, but it's not open-source and you need a personal license to use it (academic use is free).

alliecc commented 2 years ago

Hello,

Thanks for the great work. I'm a student trying to reproduce the quadratic results on another dataset. Is it possible to share the Gurobi version of code? Thanks!