search

wangch-g / lanet

MIT License
30 stars 12 forks source link

How to improve the performance of the model in matching key point #4

Open foxkw opened 1 year ago

foxkw commented 1 year ago

Hi, Thanks for your great job. I have a question :How to improve the performance of the model in matching similar key point descriptors. like buildings image

wangch-g commented 1 year ago

The duplicate patterns in local regions are hard to be matched accurately. I think you can try to use some matching algorithms, like NN+RANSAC, SuperGlue, etc., to filter the outliers, or improve the architecture of the model (attention mechanism?) to tackle this problem.